The module lifecycle stage: Generally available version

The module has requirements for installation

Deckhouse Platform installs CRDs but does not remove them when a module is disabled. If you no longer need the created CRDs, delete them.

ClusterLog

Scope: Cluster
Version: v1alpha1

Cluster-wide resource used in RBAC to control access to logs.

ClusterObservabilityAlertAdditionalLabels

Short names: coalabels

Scope: Cluster
Version: v1alpha1

Additional labels the module adds to alerts shipped with Deckhouse Platform.

If an alert already contains a label with the same key, its value is preserved.

Warning. Additional labels are applied only to alerts created using the resources labeled with heritage: deckhouse.

  • spec
    object
    • spec.additionalLabels
      object

      Required value

      Additional labels to add to each matching alert.

      If an alert already contains a label with the same key, its value is preserved.

    • spec.alertSelector
      object

      Required value

      Alerts that will receive the additional labels.

      Matching is performed against alert labels set in the rule and the alertname label with the alert name.

      • spec.alertSelector.matchExpressions
        array of objects

        Required value

        List of label selector requirements.

        All requirements must be satisfied.

        • spec.alertSelector.matchExpressions.key
          string
          Label name.
        • spec.alertSelector.matchExpressions.operator
          string

          Comparison operator:

          • In: Label exists, and its value is in values.
          • NotIn: Label does not exist, or its value is not in values.
          • Exists: Label exists.
          • DoesNotExist: Label does not exist.

          Allowed values: In, NotIn, Exists, DoesNotExist

        • spec.alertSelector.matchExpressions.values
          array of strings
          Label values for the In and NotIn operators.
      • spec.alertSelector.matchLabels
        object
        Exact key-value pairs that must be present on the alert to add labels.

ClusterObservabilityDashboard

Scope: Cluster
Version: v1alpha1

Cluster-level dashboard displayed in the Deckhouse web UI. To configure how the dashboard is displayed, use the following annotations:

  • metadata.deckhouse.io/category: Folder (category) of the dashboard in the interface. - metadata.deckhouse.io/title: Title of the dashboard in the interface. If the annotation is omitted, the module uses the title from the JSON manifest.
  • spec
    object
    • spec.definition
      string

      Required value

      Dashboard JSON manifest as a string compatible with the Grafana format.

ClusterObservabilityDiagnosticPlaybook

Short names: codp

Scope: Cluster
Version: v1alpha1

Experimental. Diagnostic playbook for firing alerts.

Tells which alerts it applies to, how the diagnosed Kubernetes object is found from the alert labels, which checks to run in addition to the standard checks of the object kind, and which deterministic findings to derive from the collected facts. The controller validates the playbook and writes the condition Valid; an invalid playbook is not run.

Only the cluster administrator can create playbooks. A playbook runs next to the built-in playbooks of the controller and cannot replace or turn one off: the names of the built-in playbooks and every name that starts with observability- are reserved, and the controller does not run a playbook with such a name (condition Valid is False with reason ReservedName).

A playbook may extend a built-in abstract playbook (spec.extends): it takes its checks and findings, and a check or a finding of its own with the id of one of them replaces it whole.

  • spec
    object
    Defines the playbook.
    • spec.abstract
      boolean
      Marks a built-in playbook that other playbooks only extend: it never matches an alert and has no match, target or baseline. Abstract playbooks may extend each other, at most three in a chain. A playbook of the cluster with abstract: true is not run: condition Valid is False with reason Invalid.
    • spec.baseline
      string
      Standard checks of the target kind: Auto runs them, None keeps only the alert context.

      Default: Auto

      Allowed values: Auto, None

    • spec.checks
      array of objects
      Checks run in addition to the baseline.
      • spec.checks.baseline
        object
        Parameters of the Baseline check: runs a built-in recipe. Its rows join the baseline: findings read them as baseline.<id>, only when the playbook has the check of the recipe; the report stores them as baseline/<id>. At most one check per recipe; needs baseline Auto and, except for VolumeUsage, ScrapeNetworkPolicies and ImageCheckLog, match.scope Cluster. Skipped with reason NotOnLinkedObject on a linked object.
        • spec.checks.baseline.patterns
          array of objects
          Lines the recipe looks for in the journals it reads, by the keys the recipe reads (EtcdMembers: staticPod, in the journal of kubelet; the other recipes read none). The journal counts the entries that match each pattern and keeps the newest one under evidence.patterns.<key> with the named groups of the expression.
          • spec.checks.baseline.patterns.expr
            string
            RE2 regular expression over the message of a journal entry; the validation of the playbook compiles it.

            Length: 1..1024

          • spec.checks.baseline.patterns.key
            string
            Key of the pattern in evidence.patterns.

            Maximum length: 63

            Pattern: ^[a-z][A-Za-z0-9]*$

        • spec.checks.baseline.recipe
          string

          Required value

          Recipe ID; the target it works on and the rows it writes. A recipe tied to a target kind ends its row NotFound with reason NoSubject on another target. ChronyTime (Node): chronyc tracking and sources from the chrony pod of the node, needs commands in containers; row node-time. CiliumAgentStatus (Pod in d8-cni-cilium): cilium-health status, cilium-dbg status --brief; row cilium-agent-status. DeckhouseModule (module of the alert label module, a ModuleRelease or a ModuleConfig): its ModulePullOverride (row module-pull-override) and, when the Module reports an error, its lines in the Deckhouse log (row controller-logs). EtcdMembers: membership and state of the etcd members of the control-plane nodes; rows control-plane-nodes, scrape-targets, etcd-membership, etcd-member-state, etcd-peer-prober, etcd-members, node-journal-kubelet, control-plane-manager. ImageCheckLog (any scope): image-availability-exporter log lines about the alert label image; row exporter-log. NodeLocalDNSUpstreams (Node): forwarding, logs and upstreams of node-local-dns and the cluster DNS backends; rows nld-forward, nld-logs, dns-backends, nld-caches, nld-upstreams. PrometheusLayout (Prometheus Pod): retention and what the entries of the data volume are, walked by the node agent; row prometheus-layout. ScrapeClientCertificate (ModuleConfig monitoring-applications or monitoring-custom): whether HTTPS scrape targets accept the client certificate of Prometheus; row scrape-client-certificate. ScrapeNetworkPolicies (Pod): whether NetworkPolicies let the main Prometheus scrape the port of the alert label instance; row scrape-network-policies. StorageTerminating (any target): storage.deckhouse.io objects with a deletion time; row storage-terminating. VolumeUsage (PersistentVolumeClaim): df, df -i and the 10 largest directories of the volume, run in the pod of the alert label pod; rows volume-df, volume-inodes, volume-du (Skipped over 400Gi).

          Allowed values: ChronyTime, CiliumAgentStatus, DeckhouseModule, EtcdMembers, ImageCheckLog, NodeLocalDNSUpstreams, PrometheusLayout, ScrapeClientCertificate, ScrapeNetworkPolicies, StorageTerminating, VolumeUsage

      • spec.checks.events
        object
        Parameters of the Events check: events of every type of the last hour of an object, grouped by message.
        • spec.checks.events.limit
          integer
          Maximum number of event groups.

          Allowed values: 1 <= X <= 50

        • spec.checks.events.object
          object
          Object whose events the check reads, in the namespace of the report.
          • spec.checks.events.object.kind
            string

            Required value

            Kind of the object.

            Allowed values: CronJob, Pod

          • spec.checks.events.object.nameTemplate
            string

            Required value

            Name of the object, exactly {{ .labels.<name> }} of an alert label, or exactly {{= <CEL> }} over the facts when reads. The computed value must be a DNS-1123 subdomain; otherwise the check is not run and ends InsufficientInput with reason Params. The object is read in the namespace of the report.

            Maximum length: 256

      • spec.checks.id
        string
        Check ID, unique within the playbook. Findings refer to it as checks.<id>; the report stores it as <playbook>/<id>.

        Maximum length: 63

        Pattern: ^[a-z0-9]([-a-z0-9]*[a-z0-9])?$

      • spec.checks.list
        object
        Parameters of the List check: lists objects of an allowlisted kind in the scope of the report, page by page.
        • spec.checks.list.kind
          string

          Required value

          Kind of the listed objects: Pod, Deployment, PersistentVolumeClaim, Node, Challenge, Order; CertificateRequest - issuer and issuance conditions, never the CSR or certificate; ClusterLoggingConfig, PodLoggingConfig (across namespaces) - cluster report only; StaticInstance - labels, phase and node, cluster report only; ModuleRelease - name, phase and status message, cluster report only.

          Allowed values: Pod, Deployment, PersistentVolumeClaim, Node, CertificateRequest, Challenge, Order, ClusterLoggingConfig, PodLoggingConfig, StaticInstance, ModuleRelease

        • spec.checks.list.labelSelector
          string
          Static label selector in Kubernetes syntax.

          Maximum length: 256

        • spec.checks.list.limit
          integer
          Maximum number of objects.

          Allowed values: 1 <= X <= 50

      • spec.checks.metrics
        object
        Parameters of the Metrics check: a named recipe of the built-in catalog over the last hour.
        • spec.checks.metrics.recipe
          string

          Required value

          Recipe ID.

          Maximum length: 63

        • spec.checks.metrics.subject
          string
          Object the recipe reads: the target, or the nodes of the pods of the target.

          Allowed values: Target, Nodes

      • spec.checks.nodeAgent
        object
        Parameters of the NodeAgent check: one operation of the node agent on the node of the target (a Node, the node of a Pod, or every control-plane node for Audit). Needs match.scope Cluster, except OOMKills and Listeners. Skipped with reason NotOnLinkedObject on a linked object and NodeAgentDisabled with alertDiagnostics.nodeAgent.enabled off. The agent reads only units and paths it allows.
        • spec.checks.nodeAgent.at
          string
          Time of Audit, RFC 3339, or exactly {{= <CEL> }} whose value is one, such as the time of the write of spec.unschedulable of the target node (baseline.target.evidence.cordonedBy.time); a computed value that is no time ends the check InsufficientInput with reason Params.

          Maximum length: 256

        • spec.checks.nodeAgent.container
          string
          Container of the target pod for OOMKills or Listeners: a DNS-1123 label or exactly {{= <CEL> }} (an invalid value ends the check InsufficientInput with reason Params, an empty one names no container). OOMKills puts the newest kill of its last terminated instance under evidence.kill; Listeners reads the sockets through its running instance.

          Maximum length: 63

        • spec.checks.nodeAgent.depth
          integer
          Directory levels below path that Disk lists; 3 when not set.

          Allowed values: 1 <= X <= 3

        • spec.checks.nodeAgent.device
          string
          Device of the file system of FilesystemErrors as the alert names it, such as /dev/sdb1, or exactly {{= <CEL> }} whose value is one, at most 256 bytes; an empty computed value names no device. Only a device of /dev outside /dev/mapper has a name the kernel lines give.

          Maximum length: 256

        • spec.checks.nodeAgent.grep
          string
          RE2 regular expression Kernel sends to the agent with the journal of the kernel, at most 256 bytes; the agent returns only the entries whose message matches it. A fixed value: the validation of the playbook compiles it.

          Length: 1..256

        • spec.checks.nodeAgent.lines
          string
          What Kernel reads in the lines it gets. ConntrackDrops: the count and the time of the newest drop of a packet for a full conntrack table, under evidence.tableFull. FilesystemErrors: the file system and I/O error lines, the newest one and the count of those that name device or its disk, and up to 3 of other devices, under evidence.fsErrors.

          Allowed values: ConntrackDrops, FilesystemErrors

        • spec.checks.nodeAgent.maxLines
          integer
          Newest entries of the window Journal asks the node agent for; 200 when not set, at most 500, the limit of the agent. The agent also returns at most 64 KiB of messages, so a journal of long lines may give fewer.

          Allowed values: 1 <= X <= 500

        • spec.checks.nodeAgent.mountpoint
          string
          Mount point the walk of Disk or DiskTree must be on, as path or exactly {{= <CEL> }}: an answer of the agent for another file system, or its refusal of the path as no mount point, ends the check InsufficientInput with reason MountNotSeen.

          Maximum length: 256

        • spec.checks.nodeAgent.operation
          string

          Required value

          Operation of the node agent. Units: states of the systemd units units. Journal: the journal of unit from 30 minutes before the alert start (at most 24 hours) to the run, the newest 200 entries or maxLines. Disk: file system use of path and its largest directories. DiskTree: file system use of path as a tree of its largest branches (up to 200 nodes, 8 levels); a walk under 30 minutes old is reused, a rerun asks for a new one. Layers: the largest writable layers of running containers. Kernel: the kernel journal in the window of Journal, filtered by grep and read as lines says. Conntrack: fill of the conntrack table and kernel counters, no connections. OOMKills: oom-kill lines of the kernel for the target Pod in the window of Journal. Listeners: listening TCP sockets of the target Pod and whether its declared ports are listened on. Audit: patches and updates of the target Node in the kube-apiserver audit log from 2 minutes before at to 1 minute after it: user, groups, user agent, source addresses, verb, time and code; the basic DKP audit policy records them only for Deckhouse service accounts. OOMKills and Listeners work with any match.scope.

          Allowed values: Units, Journal, Disk, Layers, Kernel, Conntrack, OOMKills, DiskTree, Listeners, Audit

        • spec.checks.nodeAgent.path
          string
          Path Disk and DiskTree walk: an absolute clean host path of at most 256 bytes outside /proc, /sys, /dev, /run, /var/run, /etc, /root, /home and /tmp, or exactly {{= <CEL> }} whose value is one. The agent walks /, /var/lib/containerd, /var/lib/kubelet, /var/log, /opt, and another path only when it is a mount point of ext2, ext3, ext4, xfs or btrfs.

          Maximum length: 256

        • spec.checks.nodeAgent.patterns
          array of objects
          Lines Journal looks for in the journal. The journal counts the entries that match each pattern and keeps the newest one under evidence.patterns.<key> with the named groups of the expression, up to linesBefore lines before it, and the number of lines the count is of (linesRead) and whether they are the whole window (windowRead).
          • spec.checks.nodeAgent.patterns.expr
            string
            RE2 regular expression over the message of a journal entry; the validation of the playbook compiles it.

            Length: 1..1024

          • spec.checks.nodeAgent.patterns.key
            string
            Key of the pattern in evidence.patterns.

            Maximum length: 63

            Pattern: ^[a-z][A-Za-z0-9]*$

          • spec.checks.nodeAgent.patterns.linesBefore
            integer
            Lines before the newest match the check keeps under evidence.patterns.<key>.before, oldest first: the output since the match before it, without the empty lines and the commands a shell traces (+ …).

            Allowed values: 0 <= X <= 20

        • spec.checks.nodeAgent.sortBy
          string
          Order of the directories of Disk and DiskTree and the layers of Layers, largest first: bytes (when not set) or inodes, or exactly {{= <CEL> }} whose value is one of them.

          Maximum length: 256

        • spec.checks.nodeAgent.top
          integer
          Directories Disk, or writable layers Layers, lists; 10 when not set.

          Allowed values: 1 <= X <= 20

        • spec.checks.nodeAgent.unit
          string
          Unit of Journal: kubelet, containerd, containerd-deckhouse, bashible or cloud-init, or exactly {{= <CEL> }} whose value is one of them.

          Maximum length: 256

        • spec.checks.nodeAgent.units
          array of strings
          Units of Units, each once: kubelet, containerd, containerd-deckhouse.
          • spec.checks.nodeAgent.units.Element of the array
            string

            Allowed values: kubelet, containerd, containerd-deckhouse

      • spec.checks.object
        object
        Parameters of the Object check: reads one object and stores its allowlisted projection, or finds the objects that name a Secret (bySecretName).
        • spec.checks.object.bySecretName
          object
          Finds the cert-manager Certificates in the namespace of the report whose spec.secretName is the Secret. Evidence: the first of them by name, their number (total), their names (sameSecret, up to 5) and the other Certificates of the namespace (others, up to 50). No match ends the check NotFound with reason NoMatch. The Secret itself is never read.
          • spec.checks.object.bySecretName.kind
            string

            Required value

            Kind of the objects that name the Secret.

            Allowed values: Certificate

          • spec.checks.object.bySecretName.secretNameTemplate
            string

            Required value

            Name of the Secret, a name or exactly {{ .labels.<name> }} of an alert label.

            Maximum length: 256

        • spec.checks.object.object
          object
          Object the check reads, in the namespace of the report.
          • spec.checks.object.object.kind
            string

            Required value

            Kind of the object. Pod, CronJob: in the namespace of the report. NodeGroup: cluster report only. ClusterLogDestination, IngressNginxController, NodeGroupConfiguration: cluster report only, without secrets. Module: phase, conditions, version, source and release channel of a Deckhouse module; cluster report only. AdmissionObjects: the ValidatingAdmissionPolicy with its bindings, or the webhook with its Service endpoints, named by an API server error, and whose they are (labels heritage, module); never caBundle, URL, rules or parameters; cluster report only. GatekeeperConstraint: the Gatekeeper constraint named by a Gatekeeper refusal, metadata only; cluster report only. Issuer, ClusterIssuer: the ingress class of each ACME HTTP-01 solver and labels heritage, module, never privateKeySecretRef; ClusterIssuer in a cluster report only. ClusterObservabilityDashboard, ClusterObservabilityPropagatedDashboard, ClusterObservabilityMetricsRulesGroup, ClusterObservabilityPropagatedMetricsRulesGroup: metadata only; cluster report only.

            Allowed values: Pod, CronJob, NodeGroupConfiguration, NodeGroup, ClusterLogDestination, IngressNginxController, Module, AdmissionObjects, GatekeeperConstraint, Issuer, ClusterIssuer, ClusterObservabilityDashboard, ClusterObservabilityPropagatedDashboard, ClusterObservabilityMetricsRulesGroup, ClusterObservabilityPropagatedMetricsRulesGroup

          • spec.checks.object.object.nameTemplate
            string

            Required value

            Name of the object, exactly {{ .labels.<name> }} of an alert label, or exactly {{= <CEL> }} over the facts when reads. The computed value must be a DNS-1123 subdomain; otherwise the check is not run and ends InsufficientInput with reason Params. The object is read in the namespace of the report.

            Maximum length: 256

      • spec.checks.podExec
        object
        Parameters of the PodExec check: a fixed command in a container. Runs only with alertDiagnostics.podExec.enabled.
        • spec.checks.podExec.command
          array of strings

          Required value

          Command as argv. An element {{ .labels.<name> }} is substituted as a whole argument after validation; the controller never adds a shell.
          • spec.checks.podExec.command.Element of the array
            string

            Maximum length: 256

        • spec.checks.podExec.container
          string
          Container name, or exactly {{= <CEL> }} over the facts when reads, whose value must be a DNS-1123 label (otherwise the check ends InsufficientInput with reason Params). The validation of the controller requires it.

          Maximum length: 63

        • spec.checks.podExec.limitBytes
          integer
          Limit of stdout and stderr bytes, each.

          Allowed values: 1024 <= X <= 65536

        • spec.checks.podExec.object
          object
          Object named explicitly for subject: Explicit, in the namespace of the report.
          • spec.checks.podExec.object.kind
            string

            Required value

            Kind of the object.

            Allowed values: Pod

          • spec.checks.podExec.object.nameTemplate
            string

            Required value

            Name of the object, exactly {{ .labels.<name> }} of an alert label, or exactly {{= <CEL> }} over the facts when reads. The computed value must be a DNS-1123 subdomain; otherwise the check is not run and ends InsufficientInput with reason Params. The object is read in the namespace of the report.

            Maximum length: 256

        • spec.checks.podExec.subject
          string

          Required value

          Pod the command runs in: the target pod (the first representative pod for a workload target), or the pod of object.

          Allowed values: Target, Explicit

        • spec.checks.podExec.timeoutSeconds
          integer
          Command timeout in seconds.

          Allowed values: 1 <= X <= 60

      • spec.checks.podLogs
        object
        Parameters of the PodLogs check: container logs from the kubelet, at most alertDiagnostics.logs.maxBytesPerCheck bytes per container.
        • spec.checks.podLogs.container
          string
          Container name, or exactly {{= <CEL> }} over the facts when reads, whose value must be a DNS-1123 label (otherwise the check ends InsufficientInput with reason Params). By default the failing container, otherwise the first one.

          Maximum length: 63

        • spec.checks.podLogs.grep
          string
          RE2 regular expression the kept lines must match.

          Maximum length: 128

        • spec.checks.podLogs.object
          object
          Object named explicitly for subject: Explicit, in the namespace of the report.
          • spec.checks.podLogs.object.kind
            string

            Required value

            Kind of the object.

            Allowed values: Pod

          • spec.checks.podLogs.object.nameTemplate
            string

            Required value

            Name of the object, exactly {{ .labels.<name> }} of an alert label, or exactly {{= <CEL> }} over the facts when reads. The computed value must be a DNS-1123 subdomain; otherwise the check is not run and ends InsufficientInput with reason Params. The object is read in the namespace of the report.

            Maximum length: 256

        • spec.checks.podLogs.previous
          boolean
          Read the previous instance of the container.
        • spec.checks.podLogs.subject
          string

          Required value

          Pods the check reads: the target pod (the representative pods for a workload target), the representative pods of the target, or the pod of object.

          Allowed values: Target, RepresentativePods, Explicit

        • spec.checks.podLogs.tailLines
          integer
          Number of last lines.

          Allowed values: 1 <= X <= 200

      • spec.checks.scrapeTargets
        object
        Parameters of the ScrapeTargets check: label values a Prometheus scrape target must have to be read; an unset field matches any value. Targets are read with the permissions of the report, so a namespace report sees only its own targets. Each value is a label value or exactly {{= <CEL> }}; a computed value that is empty or over 256 bytes ends the check InsufficientInput with reason Params.
        • spec.checks.scrapeTargets.job
          string

          Required value

          Value of the label job of the target.

          Length: 1..256

        • spec.checks.scrapeTargets.namespace
          string
          Value of the label namespace of the target.

          Maximum length: 256

        • spec.checks.scrapeTargets.pod
          string
          Value of the label pod of the target.

          Maximum length: 256

      • spec.checks.serviceBackends
        object
        Parameters of the ServiceBackends check: endpoints of a Service and the pods behind them. Without parameters, the Service of the target route, or for a Pod target the Services that select the pod.
        • spec.checks.serviceBackends.serviceFromLabel
          string
          Alert label that holds the Service name.

          Maximum length: 63

      • spec.checks.title
        string
        Name of the check for a reader of the report, in sentence case and 1 to 4 words, such as CronJob of the Job: status.checks[].title of its row. Plain text without a template. Without it the report names the check by its type and subject (Pod logs, Metrics: memory usage). A check of type Baseline that runs has no row of its own: the rows of its recipe are named by the controller.

        Length: 1..64

      • spec.checks.type
        string
        Check type.

        Allowed values: Object, List, Events, PodLogs, PodExec, Metrics, ServiceBackends, ScrapeTargets, Baseline, NodeAgent

      • spec.checks.when
        string
        CEL condition over baseline.<id>, checks.<id>, values.baseline.<check>.<value>, alert and target; the check runs only when it is true, after the checks it reads. A false condition records the check Skipped with reason ConditionFalse (this does not make the collection Partial); a condition that fails to evaluate records it InsufficientInput with reason Params.

        Length: 1..8192

    • spec.component
      string
      Platform component the playbook diagnoses. Its causes are about this component and the components it covers, so their own failure during the alert does not withdraw such a cause. When a withdrawn cause blames an unready component, it names the report of a firing cluster alert whose playbook has this component. Adds no reads. Not allowed in an abstract playbook.

      Allowed values: kube-apiserver, etcd, kube-dns, node-local-dns, cni-agent, node, deckhouse, ingress-controller, prometheus, cert-manager, kruise-controller, grafana, prometheus-operator, extended-monitoring-exporter, prometheus-longterm, image-availability-exporter, trickster, kube-state-metrics, cluster-autoscaler, machine-controller-manager, bashible-apiserver, caps-controller-manager, cloud-data-discoverer, upmeter, terraform-state-exporter, snapshot-controller, dex, gatekeeper, documentation, stronghold, sds-local-volume, safe-agent-updater, node-exporter, upmeter-agent, log-shipper, cilium-operator, user-authn-controller, dexclient-controller, permission-browser-apiserver, user-authz-controller, user-authz-webhook, dop-tempo, dop-loki, registry-controller, registry-storage, registry-agent, dop-ceph, dop-flow-collector, dop-mimir, dop-alertgate, dop-assistant, dop-auth, dop-backend, dop-reconciler, dop-rsm-engine, dop-metrics-collector, dop-usage-collector, control-plane-manager, node-controller, multitenancy-manager, x509-certificate-exporter, caphc-controller-manager, capi-controller-manager

    • spec.extends
      string
      Name of a built-in abstract playbook whose checks and findings this playbook inherits; a playbook of the cluster cannot be a parent. A check or finding with the id of one of the parent replaces it in place, the others follow. match, target and baseline are not inherited. At most 64 findings after the merge. A wrong parent sets Valid False with reason ExtendsNotFound, ExtendsNotAbstract or ExtendsInvalid.

      Maximum length: 253

      Pattern: ^[a-z0-9]([-a-z0-9]*[a-z0-9])?(\.[a-z0-9]([-a-z0-9]*[a-z0-9])?)*$

    • spec.findings
      array of objects
      Findings of the playbook.
      • spec.findings.actions
        array of strings
        What to do, Markdown templates.
        • spec.findings.actions.Element of the array
          string

          Maximum length: 4096

      • spec.findings.causedBy
        array of strings
        Findings this one follows from: IDs of findings of the same playbook or IDs of built-in findings of the controller (baseline/<id>). In a report the finding names the causes that hold in the same run and comes after the findings that stand on their own.
        • spec.findings.causedBy.Element of the array
          string

          Maximum length: 72

          Pattern: ^(baseline/)?[a-z0-9]([-a-z0-9]*[a-z0-9])?$

      • spec.findings.evidence
        array of strings
        Additions to the checks the finding is based on, baseline.<id> or <check id>: a check the finding cites without reading it, or one it reads that the report cites first. The checks when and the texts read are added after them. The finding is evaluated as soon as all of them are collected.
        • spec.findings.evidence.Element of the array
          string

          Maximum length: 80

      • spec.findings.explanation
        string
        Markdown without HTML; template with {{ .labels.<name> }} and {{= <CEL> }}.

        Maximum length: 16384

      • spec.findings.facts
        array of objects
        The values that prove the finding, which a reader sees next to it in this order: each a label and a value. A fact whose value renders empty is left out of the report, and so is one whose expression fails, which the condition FindingsEvaluated reports. The checks the values read are added to the evidence of the finding after those of when and of the texts.
        • spec.findings.facts.highlight
          boolean
          The fact carries the problem itself, such as the text of an error, the reason of a refusal or the message of a component: the report keeps the mark in facts[].highlight, and the console shows the fact first, above the other facts of the finding. At most 2 facts of a finding are marked.
        • spec.findings.facts.inPrimaryEvidence
          boolean
          The fact only repeats what the primary evidence of the finding shows, such as the largest directory of a tree: the report keeps it and marks it inPrimaryEvidence, and the console leaves it out where it draws that evidence. Set only on a fact of a finding with primaryEvidence.
        • spec.findings.facts.label
          string
          A short noun phrase a reader of a table understands without the title, such as Memory limit. Plain text without a template.

          Length: 1..64

        • spec.findings.facts.value
          string
          The value with its unit and nothing else, such as 512 MiB: a template with {{ .labels.<name> }}, {{ .target.<field> }} and {{= <CEL> }}, as a title. It renders into one line cut to 256 bytes.

          Length: 1..1024

      • spec.findings.hypothesis
        string
        Platform hypothesis the finding confirms, in checking order: Infrastructure (node, network, storage, API, registry), ClientSetting (a setting the user can change: a ModuleConfig or module resource), Dependency (another platform component). Only on a target of class Platform: the earliest step among the causes wins, later ones go to status.platformNotes. Default: ClientSetting for ConfigurationCause, else Infrastructure; none for Symptom.

        Allowed values: Infrastructure, ClientSetting, Dependency

      • spec.findings.id
        string
        Finding ID, unique within the playbook. The report stores it as <playbook>/<id>.

        Maximum length: 63

        Pattern: ^[a-z0-9]([-a-z0-9]*[a-z0-9])?$

      • spec.findings.platformIndependent
        boolean
        Marks a cause read from the object itself, independent of the API server, DNS and the platform network (e.g. an OOM kill, the HTTP status of a probe). A basic component not ready before the alert does not withdraw it and is noted in status.platformNotes. Only for role Cause with severity Warning or Critical. A finding that replaces one of the parent sets its own value.
      • spec.findings.primaryEvidence
        string
        The check whose evidence proves the finding, which the console draws with it, such as the tree of a disk: baseline.<id> or <check id>, as an entry of evidence. It is among the evidence of the finding, an entry of evidence or a check its expressions read, else the playbook is invalid. The report gives its id in findings[].primaryEvidence when the run holds the check.

        Length: 1..80

      • spec.findings.references
        array of objects
        Up to 3 links of the finding to a Deckhouse module, in this order. The console builds the URLs to the module settings and documentation; a playbook never gives a URL. module, setting, resourceKind, resourceVersion and anchor are templates, as in a title. A reference whose module renders empty, or whose expression fails, is dropped.
        • spec.findings.references.anchor
          string
          ModuleDocs only: the anchor on the page, as in the published documentation.

          Maximum length: 1024

        • spec.findings.references.kind
          string
          ModuleSetting: a key of the settings of the module (its ModuleConfig). ModuleResource: a kind of a resource of the module. ModuleDocs: a page of the documentation of the module.

          Allowed values: ModuleSetting, ModuleResource, ModuleDocs

        • spec.findings.references.module
          string
          The module by its name in Module and ModuleConfig, such as upmeter or ingress-nginx; global for the global ModuleConfig, with the kind ModuleSetting only. A module without documentation (monitoring-deckhouse) is not named.

          Length: 1..1024

        • spec.findings.references.page
          string
          ModuleDocs only: the page of the documentation of the module.

          Allowed values: Overview, Configuration, Resources, Usage, FAQ, Examples

        • spec.findings.references.resourceKind
          string
          ModuleResource only: the kind of a resource of the module, such as NodeGroup.

          Maximum length: 1024

        • spec.findings.references.resourceVersion
          string
          ModuleResource only: the version of the API of the resource, such as v1. Without it the controller takes the version of the object of the finding when it is of that kind, else the version it reads the kind with.

          Maximum length: 1024

        • spec.findings.references.setting
          string
          ModuleSetting only: the dot-separated path of the key in the settings of the module, without array indexes, such as smokeMini.storageClass. Never a value.

          Maximum length: 1024

      • spec.findings.role
        string
        What the finding says about the cause of the alert: Cause - a behaviour cause, turned into a symptom when a basic component on the path of the target was not ready during the alert; ConfigurationCause - a cause in configuration or object state, such as a ResourceQuota or NetworkPolicy, never withdrawn; Symptom - restates what the alert or object shows. A headline from a symptom starts with Cause not established:.

        Default: Cause

        Allowed values: Cause, ConfigurationCause, Symptom

      • spec.findings.severity
        string
        Severity.

        Allowed values: Info, Warning, Critical

      • spec.findings.title
        string
        Title template, one line. The rendered title is cut to 200 bytes.

        Length: 1..1024

      • spec.findings.unless
        array of strings
        IDs of other findings of the playbook, after the merge with extends: this finding is not written when any of them is written in the same evaluation for the same target or linked object. A named finding whose condition fails silences nothing. Use it for a fallback; for a consequence that holds together with its cause, use causedBy.
        • spec.findings.unless.Element of the array
          string

          Maximum length: 63

          Pattern: ^[a-z0-9]([-a-z0-9]*[a-z0-9])?$

      • spec.findings.when
        string
        CEL expression over baseline.<id>, checks.<id>, values.baseline.<check>.<value> and alert; the finding holds when it returns true. Required, except in a finding of a playbook with extends that replaces a finding of the parent by its id: without when it takes the condition of the finding it replaces, and nothing else of it.

        Length: 1..8192

    • spec.let
      array of objects
      Named CEL expressions, read as let.<name> in the conditions and {{= <CEL> }} expressions of this playbook and of the playbooks that extend it, as if written in place. With the names replaced, an expression has at most 16384 characters. A name must not repeat a name of the parent in extends, and in a playbook that is not abstract every name must be read.
      • spec.let.expr
        string
        CEL expression over the variables of a finding (baseline, checks, values, alert, target) that compiles on its own and reads no let. The checks and the facts it reads are checked in each expression that reads it, against the target of that playbook.

        Length: 1..2048

      • spec.let.name
        string
        Name the expressions read as let.<name>; not in, true, false or null, which CEL reads as words of its own.

        Maximum length: 32

        Pattern: ^[a-z][a-zA-Z0-9]*$

    • spec.match
      object
      Alerts the playbook applies to, and with object the linked objects of a run.
      • spec.match.alertNames
        array of strings
        Names of the alerts; may be omitted with object, then the playbook never matches an alert by its name.
        • spec.match.alertNames.Element of the array
          string

          Length: 1..253

      • spec.match.labelSelector
        object
        Selector over the alert labels.
        • spec.match.labelSelector.matchExpressions
          array of objects
          Expressions over labels; all of them must hold.
          • spec.match.labelSelector.matchExpressions.key
            string
            Label name.

            Maximum length: 63

          • spec.match.labelSelector.matchExpressions.operator
            string
            Operator.

            Allowed values: In, Exists, DoesNotExist

          • spec.match.labelSelector.matchExpressions.values
            array of strings
            Values for In.
            • spec.match.labelSelector.matchExpressions.values.Element of the array
              string

              Maximum length: 256

        • spec.match.labelSelector.matchLabels
          object
          Labels and their values.
      • spec.match.object
        object
        Applies the playbook also to a linked object of a run of kind target.kind that got the rest of its baseline, with the labels namespace and target.nameFromLabel set to the object; {} matches any such object. Needs target.nameFromLabel.
        • spec.match.object.labelSelector
          object
          Selector over the labels of the object itself.
          • spec.match.object.labelSelector.matchLabels
            object
            Labels and their values.
        • spec.match.object.ports
          array of integers
          Ports, one of which must be the port of the address that led to the object or a port a container of its pods declares.
          • spec.match.object.ports.Element of the array
            integer

            Allowed values: 1 <= X <= 65535

      • spec.match.requiredLabels
        array of strings
        Alert labels the target needs; an alert without one of them does not address the target, and the playbook does not take part in the run.
        • spec.match.requiredLabels.Element of the array
          string

          Maximum length: 63

      • spec.match.scope
        string

        Required value

        Alerts to match: Namespaced - ObservabilityAlert; Cluster - ClusterObservabilityAlert; Any - both, for a rule whose alert changed scope between DKP versions. The report has the scope of the alert. Any has the limits of Namespaced: the target is in the namespace of the alert (or its label namespace); no target.namespace, target.namespaceFromLabel, target.selector, kind Node or checks that need Cluster.

        Allowed values: Namespaced, Cluster, Any

      • spec.match.targetClass
        string
        Class of target the playbook applies to: User - a user workload; Platform - a Deckhouse object (label heritage: deckhouse on it or its namespace, a d8-* or kube-system namespace, a Node, a cluster resource of a Deckhouse API group); Any - either. A playbook of another class does not run, and condition FindingsEvaluated names it. With object, the class of the linked object counts.

        Default: User

        Allowed values: User, Platform, Any

    • spec.target
      object
      How the diagnosed object is found from the alert labels.
      • spec.target.apiVersion
        string

        Required value

        API version of the object.

        Maximum length: 128

      • spec.target.followBackend
        boolean
        Only with kind Ingress: also diagnose the backend (the Deployment, StatefulSet or DaemonSet behind the Service of the route): it is linked in status.run.linked with relation Backend, its pods and logs are read, and its built-in causes, such as backend-errors, become causes of the Warning and Critical findings. For alerts about backend answers: 5xx, latency, unavailable backends.
      • spec.target.kind
        string

        Required value

        Kind of the object.

        Allowed values: Pod, Deployment, StatefulSet, DaemonSet, Job, CronJob, Service, Ingress, PersistentVolumeClaim, Node, DeckhouseRelease, ModuleRelease, NodeGroup, ModuleConfig, Certificate

      • spec.target.nameFromLabel
        string
        Alert label that holds the object name.

        Maximum length: 63

      • spec.target.nameTemplate
        string
        Template that builds the object name from alert labels.

        Maximum length: 256

      • spec.target.namespace
        string
        Namespace of the object; only with match.scope Cluster.

        Maximum length: 63

      • spec.target.namespaceFromLabel
        string
        Alert label that holds the namespace of the object; only with match.scope Cluster.

        Maximum length: 63

      • spec.target.secretNameFromLabel
        string
        Alert label that holds a Secret name; only with kind Ingress. The target is the Ingress in the namespace of the alert whose spec.tls[].secretName is that Secret: of several, the first by name whose hosts include the alert label host, else the first by name. All matches, up to 50, are listed in the baseline check peers. No match leaves the target unresolved. The Secret is never read.

        Maximum length: 63

      • spec.target.selector
        object
        Selects the object by labels in a fixed namespace; for match.scope Cluster and kinds Pod, Deployment, StatefulSet, DaemonSet only. The pod on the node of the alert label node is taken first; all matching objects, up to 50, are listed in the baseline check peers.
        • spec.target.selector.matchLabels
          object
          Labels and their values.
  • status
    object
    Written by the controller.
    • status.checkOrder
      array of objects
      Order the checks of the valid playbook run in. A check that reads checks.<id> in when or in a {{= }} parameter runs in a later stage than that check; the checks of one stage run together, in the order of the merged playbook. Absent for an invalid playbook and for one without checks.
      • status.checkOrder.id
        string
        Check ID.

        Maximum length: 63

      • status.checkOrder.stage
        integer
        1 for a check that reads no check of the playbook, otherwise one more than the latest stage of the checks it reads.

        Allowed values: 1 <= X <= 40

    • status.conditions
      array of objects
      Conditions.
      • status.conditions.lastTransitionTime
        string
        Time of the last change of the status.
      • status.conditions.message
        string
        Message for a human.

        Maximum length: 32768

      • status.conditions.observedGeneration
        integer
        Generation the condition describes.

        Allowed values: 0 <= X

      • status.conditions.reason
        string
        Machine-readable reason.

        Maximum length: 1024

      • status.conditions.status
        string
        Status.

        Allowed values: True, False, Unknown

      • status.conditions.type
        string
        Type.

        Maximum length: 316

    • status.extends
      object
      The built-in playbook the playbook extends, as the controller merged them, and where each check and finding of the merged playbook comes from. Absent for a playbook that extends none or whose parent is not found, not abstract or invalid.
      • status.extends.checks
        array of objects
        Checks of the merged playbook in their order.
        • status.extends.checks.id
          string
          Check ID.

          Maximum length: 63

        • status.extends.checks.origin
          string
          Common: a check of the parent; Replaced: a check of the playbook with the id of a check of the parent, in its place; Own: a check of the playbook the parent does not have.

          Allowed values: Common, Replaced, Own

      • status.extends.findings
        array of objects
        Findings of the merged playbook in their order.
        • status.extends.findings.id
          string
          Finding ID.

          Maximum length: 63

        • status.extends.findings.origin
          string
          Common: a finding of the parent; Replaced: a finding of the playbook with the id of a finding of the parent, in its place; Own: a finding of the playbook the parent does not have.

          Allowed values: Common, Replaced, Own

      • status.extends.name
        string

        Required value

        Name of the parent.

        Maximum length: 253

      • status.extends.specHash
        string

        Required value

        sha256:<hex> of the spec of the parent the playbook was merged with; it changes when an update of the module changes the parent.

        Maximum length: 71

    • status.observedGeneration
      integer
      Generation the conditions describe.

ClusterObservabilityDiagnosticReport

Short names: codr

Scope: Cluster
Version: v1alpha1

Experimental. Diagnostic report of a cluster alert (ClusterObservabilityAlert).

The controller creates the report when the alert fires: it finds the object of the alert, collects facts and evaluates the findings of the playbooks. A user with edit rights can create a report with only spec.alert.kind and spec.alert.name set to run the diagnostics manually, or request a new run of an existing report with the annotation observability.deckhouse.io/rerun.

Only the controller writes status.

  • spec
    object
    Defines the report. Immutable once the controller filled the alert snapshot.
    • spec.alert
      object

      Required value

      Reference to the alert object and the snapshot of the alert. A user creating a report sets only kind, name and namespace; the controller fills the snapshot.
      • spec.alert.alertName
        string
        Name of the alert.

        Maximum length: 253

      • spec.alert.annotations
        object
        Allowlisted annotations of the alert, each at most 1 KiB and at most 4 KiB together; the controller cuts a longer annotation and ends it with [truncated].
      • spec.alert.createdBy
        string
        Who created the rule of the alert.

        Maximum length: 253

      • spec.alert.fingerprint
        string
        Fingerprint of the alert.

        Maximum length: 64

      • spec.alert.kind
        string

        Required value

        Kind of the alert object.

        Allowed values: ClusterObservabilityAlert

      • spec.alert.labels
        object
        Labels of the alert, at most 2 KiB together with their names.
      • spec.alert.name
        string

        Required value

        Name of the alert object.

        Length: 1..253

      • spec.alert.namespace
        string
        Namespace of the alert object; for ObservabilityDiagnosticReport the namespace of the report.

        Maximum length: 63

      • spec.alert.rule
        object
        Rule context of the alert.
        • spec.alert.rule.description
          string
          Description template of the rule when it differs from the annotation.

          Maximum length: 2048

        • spec.alert.rule.expr
          string
          Expression of the rule.

          Maximum length: 2048

        • spec.alert.rule.for
          string
          for of the rule.

          Maximum length: 32

        • spec.alert.rule.keepFiringFor
          string
          keepFiringFor of the rule.

          Maximum length: 32

        • spec.alert.rule.summary
          string
          Summary template of the rule when it differs from the annotation.

          Maximum length: 1024

      • spec.alert.ruleGroupName
        string
        Rule group of the alert.

        Maximum length: 253

      • spec.alert.severity
        string
        Severity level of the alert.

        Maximum length: 16

      • spec.alert.startsAt
        string
        Time the alert started firing.
  • status
    object
    Written by the controller only.
    • status.alertState
      object
      State of the alert.
      • status.alertState.lostAt
        string
        When the report became Lost: once the alert object has stayed missing for the grace of the controller (3 minutes by default) without resolving.
      • status.alertState.resolvedAt
        string
        When the alert resolved.
      • status.alertState.state
        string
        State.

        Allowed values: Firing, Resolved, Lost

    • status.checks
      array of objects
      Results of the checks of the current run.
      • status.checks.coverage
        object
        How many objects of the population were checked in detail.
        • status.checks.coverage.checked
          integer

          Required value

          Checked in detail.

          Allowed values: 0 <= X

        • status.checks.coverage.total
          integer

          Required value

          Population.

          Allowed values: 0 <= X

      • status.checks.evidence
        object
        Allowlisted projection of the read data; the shape depends on the check type.
      • status.checks.evidenceKind
        string
        Name of the schema of evidence, one per shape (Events, PodLogs, Metrics, Deployment, WorkloadPods, Omitted when the budget of the run replaced the evidence). Absent on a row without evidence and in a report written before the field existed.

        Maximum length: 64

        Pattern: ^[A-Z][A-Za-z0-9]*$

      • status.checks.hidden
        boolean
        The row is not listed to a reader - a repeat with sameAs, or the snapshot of the rule baseline/alert-context.
      • status.checks.id
        string
        baseline/<check>, baseline/<kind>-<name>/<check> with several targets (baseline/<kind>-<namespace>-<name>/<check> when another target has its kind and name, with -<hash> after the name when two targets would share a prefix), baseline/<n>-<kind>-<name>/<check> of a linked object, or <playbook>/<check>.

        Maximum length: 320

      • status.checks.objectRef
        object
        Object the check read.
        • status.checks.objectRef.apiVersion
          string
          API version.

          Maximum length: 128

        • status.checks.objectRef.kind
          string

          Required value

          Kind.

          Maximum length: 63

        • status.checks.objectRef.name
          string

          Required value

          Name.

          Maximum length: 253

        • status.checks.objectRef.namespace
          string
          Namespace.

          Maximum length: 63

        • status.checks.objectRef.resourceVersion
          string
          Resource version at the time of reading.

          Maximum length: 64

        • status.checks.objectRef.uid
          string
          UID.

          Maximum length: 64

      • status.checks.observedAt
        string
        Time the fact was read.
      • status.checks.origin
        string
        Who planned the check.

        Allowed values: Baseline, Playbook

      • status.checks.outcome
        string
        Outcome. Anything other than Ok and ProblemFound means “unknown”, not “no problem”.

        Allowed values: Ok, ProblemFound, NotFound, InsufficientInput, Forbidden, Timeout, Error, Skipped

      • status.checks.reason
        string
        Stable CamelCase reason of the outcome.

        Maximum length: 63

      • status.checks.result
        string
        Verdict of the check: Problem - found a deviation, even in a partial read; Ok - a check with its own verdict found nothing; Partial - no deviation, but only part was read or the evidence was cut; NoData - nothing to judge (outcome InsufficientInput, NotFound, Skipped, or no metric sample in the window); Failed - no answer (outcome Error, Forbidden, Timeout); Context - data without a verdict. Absent in older reports.

        Allowed values: Problem, Ok, Partial, NoData, Failed, Context

      • status.checks.sameAs
        string
        ID of an earlier check of the run with the same canonical key; such a check stores no evidence.

        Maximum length: 320

      • status.checks.summary
        string
        One line for a human, in English.

        Maximum length: 300

      • status.checks.targetRef
        object
        Target of the run (an entry of status.run.targets) or linked object (an entry of status.run.linked) the row belongs to. Absent on the rows of the run as a whole (baseline/alert-context, baseline/linked-alerts) and in a report written before the field existed. objectRef stays the object the check read.
        • status.checks.targetRef.apiVersion
          string
          API version.

          Maximum length: 128

        • status.checks.targetRef.kind
          string

          Required value

          Kind.

          Maximum length: 63

        • status.checks.targetRef.name
          string

          Required value

          Name.

          Maximum length: 253

        • status.checks.targetRef.namespace
          string
          Namespace.

          Maximum length: 63

        • status.checks.targetRef.resourceVersion
          string
          Resource version at the time of reading.

          Maximum length: 64

        • status.checks.targetRef.uid
          string
          UID.

          Maximum length: 64

      • status.checks.title
        string
        Name of the check for a reader, in English, 1 to 4 words: of the table of the controller for a row of the baseline, spec.checks[].title of the playbook for a check of a playbook, else built from its type and subject. The controller writes it on every row; absent in a report written before the field existed.

        Maximum length: 64

      • status.checks.truncated
        boolean
        Evidence was cut by the limits; the summary names what was removed.
      • status.checks.type
        string
        Check type; NodeAgent is a call of the node agent of the node of a Node target, ScrapeTargets reads the Prometheus scrape targets through label-enforcer, Baseline is a check of a recipe of a playbook that did not run (its condition is false, a parameter failed or the object is linked), whose recipe wrote no row.

        Allowed values: Object, Related, List, Events, PodLogs, PodExec, Metrics, RelatedAlerts, WorkloadContext, ServiceBackends, AlertContext, NodeAgent, ScrapeTargets, Baseline

    • status.collection
      object
      Fact collection of the current run.
      • status.collection.checksDone
        integer
        Finished checks.

        Allowed values: 0 <= X

      • status.collection.checksTotal
        integer
        Planned checks.

        Allowed values: 0 <= X

      • status.collection.completedAt
        string
        End.
      • status.collection.message
        string
        Message for a human.

        Maximum length: 1024

      • status.collection.phase
        string
        Phase.

        Allowed values: Pending, Running, Completed, Partial, Failed

      • status.collection.reason
        string
        Reason of a failed collection.

        Maximum length: 63

      • status.collection.startedAt
        string
        Start.
    • status.conditions
      array of objects
      Conditions.
      • status.conditions.lastTransitionTime
        string
        Time of the last change of the status.
      • status.conditions.message
        string
        Message for a human.

        Maximum length: 32768

      • status.conditions.observedGeneration
        integer
        Generation the condition describes.

        Allowed values: 0 <= X

      • status.conditions.reason
        string
        Machine-readable reason.

        Maximum length: 1024

      • status.conditions.status
        string
        Status.

        Allowed values: True, False, Unknown

      • status.conditions.type
        string
        Type.

        Maximum length: 316

    • status.environment
      object
      Snapshot of versions.
      • status.environment.dkpVersion
        string
        Deckhouse version.

        Maximum length: 64

      • status.environment.kubernetesVersion
        string
        Kubernetes version.

        Maximum length: 64

      • status.environment.modules
        array of objects
        Modules the playbooks name.
        • status.environment.modules.enabled
          boolean
          Whether the module is enabled.
        • status.environment.modules.name
          string
          Name.

          Maximum length: 63

      • status.environment.observabilityVersion
        string
        Version of the module.

        Maximum length: 64

    • status.findings
      array of objects
      Findings of the current run.
      • status.findings.actions
        array of strings
        What to do, in Markdown.
        • status.findings.actions.Element of the array
          string

          Maximum length: 1024

      • status.findings.causedBy
        array of strings
        IDs of the findings of this run the finding follows from. Empty for a finding that stands on its own; such findings come first.
        • status.findings.causedBy.Element of the array
          string

          Maximum length: 128

      • status.findings.demotion
        string
        Why a behaviour cause was taken back: the sentence on the basic component that was not ready in the window of the alert. status.summary.causeGap repeats it when the finding gives the headline.

        Maximum length: 512

      • status.findings.evidenceRefs
        array of strings
        IDs of the checks the finding is based on.
        • status.findings.evidenceRefs.Element of the array
          string

          Maximum length: 320

      • status.findings.explanation
        string
        Why, in Markdown.

        Maximum length: 4096

      • status.findings.facts
        array of objects
        The values that prove a finding, in the order of its playbook, or of the file of the texts of the controller for a built-in finding: a fact whose value rendered empty, or whose expression failed, is not here. A built-in finding may carry facts too.
        • status.findings.facts.checkId
          string
          ID of the first check the expressions of the value read that the run holds, in the form of evidenceRefs: baseline/<check> with the prefix of the target or of the linked object, or <playbook>/<check>. Empty for a value that reads no check or none the run holds. For a fact of a built-in finding, the check its value was read from, else the first evidenceRefs of the finding.

          Maximum length: 320

        • status.findings.facts.highlight
          boolean
          The fact carries the problem itself, such as the text of an error, the reason of a refusal or the message of a component: its playbook, or the file of the texts of the controller for a built-in finding, marks it. A reader shows it first. At most 2 facts of a finding are marked; a fact whose value rendered empty is not here, with its mark.
        • status.findings.facts.inPrimaryEvidence
          boolean
          The fact only repeats what primaryEvidence of the finding shows, such as the largest directory of a tree: a reader that draws that evidence with the finding may leave the fact out. Set only when the finding has primaryEvidence.
        • status.findings.facts.label
          string
          What the value is, as the playbook names it.

          Maximum length: 64

        • status.findings.facts.value
          string
          The value in one line.

          Maximum length: 256

      • status.findings.id
        string
        <playbook>/<finding>, baseline/<finding> for a built-in finding of the controller, or baseline/<n>-<kind>-<name>/<finding> for a built-in finding of a linked object.

        Maximum length: 128

      • status.findings.objectRef
        object
        Object the finding is about: the target of its playbook run or of the prefix of its id (an entry of status.run.targets), or the linked object of its playbook run or of its prefix (an entry of status.run.linked). Absent on a built-in finding about no single object and in a report written before the field existed.
        • status.findings.objectRef.apiVersion
          string
          API version.

          Maximum length: 128

        • status.findings.objectRef.kind
          string

          Required value

          Kind.

          Maximum length: 63

        • status.findings.objectRef.name
          string

          Required value

          Name.

          Maximum length: 253

        • status.findings.objectRef.namespace
          string
          Namespace.

          Maximum length: 63

        • status.findings.objectRef.resourceVersion
          string
          Resource version at the time of reading.

          Maximum length: 64

        • status.findings.objectRef.uid
          string
          UID.

          Maximum length: 64

      • status.findings.primaryEvidence
        string
        ID of the check whose evidence proves the finding, in the form of evidenceRefs: the primaryEvidence of its playbook, set only when the run holds that check. The console draws that evidence with the finding. Absent for a built-in finding and when the playbook names none.

        Maximum length: 320

      • status.findings.references
        array of objects
        Up to 3 links of the finding to a Deckhouse module, in the order of its playbook. The console builds the URL: ModuleSetting - the module settings and the parameter in its documentation; ModuleResource - the resource in the module documentation; ModuleDocs - a documentation page. Never a URL or a setting value; a console that does not know a kind skips it. Absent in older reports.
        • status.findings.references.anchor
          string
          ModuleDocs only: the anchor on the page, as in the published documentation.

          Maximum length: 128

          Pattern: ^[\p{L}\p{N}_][\p{L}\p{N}_.-]*$

        • status.findings.references.kind
          string
          What the reference points to.

          Allowed values: ModuleSetting, ModuleResource, ModuleDocs

        • status.findings.references.module
          string
          The module by its name in Module and ModuleConfig; global for the global ModuleConfig.

          Maximum length: 64

          Pattern: ^[a-z0-9]([-a-z0-9]*[a-z0-9])?$

        • status.findings.references.page
          string
          ModuleDocs only: the page of the documentation of the module.

          Allowed values: Overview, Configuration, Resources, Usage, FAQ, Examples

        • status.findings.references.resourceKind
          string
          ModuleResource only: the kind of a resource of the module.

          Maximum length: 63

          Pattern: ^[A-Z][A-Za-z0-9]*$

        • status.findings.references.resourceVersion
          string
          ModuleResource only: the version of the API of the resource, such as v1; absent when it is not known.

          Maximum length: 32

          Pattern: ^v\d+((alpha|beta)\d+)?$

        • status.findings.references.setting
          string
          ModuleSetting only: the dot-separated path of the key in the settings of the module, without array indexes. Never a value.

          Maximum length: 256

          Pattern: ^[A-Za-z_][A-Za-z0-9_-]*(\.[A-Za-z_][A-Za-z0-9_-]*)*$

      • status.findings.role
        string
        What the finding says about the cause of the alert: Cause, ConfigurationCause or Symptom as set in the playbook (a withdrawn behaviour cause is Symptom); Fact for a built-in fact; ObjectProblem for a would-be cause found by a fallback playbook: a problem of the object from the alert labels, not the cause of the alert. symptom is true exactly for Symptom.

        Allowed values: Cause, ConfigurationCause, Symptom, Fact, ObjectProblem

      • status.findings.severity
        string
        Severity.

        Allowed values: Info, Warning, Critical

      • status.findings.symptom
        boolean
        The finding only restates what the alert or the object shows and does not establish the root cause by itself. The marker does not change causedBy.
      • status.findings.title
        string
        Title, one line.

        Maximum length: 200

    • status.history
      array of objects
      Up to five past runs.
      • status.history.completedAt
        string
        End.
      • status.history.findings
        array of objects
        Findings.
        • status.history.findings.evidenceRefs
          array of strings
          IDs of the checks.
          • status.history.findings.evidenceRefs.Element of the array
            string

            Maximum length: 320

        • status.history.findings.id
          string
          Finding ID.

          Maximum length: 128

        • status.history.findings.severity
          string
          Severity.

          Allowed values: Info, Warning, Critical

        • status.history.findings.title
          string
          Title.

          Maximum length: 200

      • status.history.headline
        string
        Headline.

        Maximum length: 300

      • status.history.id
        string
        Run ID.

        Maximum length: 64

      • status.history.startedAt
        string
        Start.
      • status.history.trigger
        string
        Trigger.

        Allowed values: Alert, Manual

    • status.platformNotes
      array of objects
      One-line notes under the findings. For a target of class Platform, first the findings of a later hypothesis step that were not taken as the cause, one per finding. Then the basic components on the path of a target that were not ready during the alert but did not withdraw a cause, because their failure began after the alert or at an unknown time; one per component and target.
      • status.platformNotes.checkId
        string
        ID of the platform-state check the note was read from; for a note of a hypothesis, the first evidence of the finding.

        Maximum length: 320

      • status.platformNotes.component
        string
        Name of the component; in a project report only the class of the failure. A note of a hypothesis names the module of the target, else the component of the playbook of the finding, else platform.

        Maximum length: 63

      • status.platformNotes.message
        string
        One sentence: what was not ready, the source it was read from, and that its failure started after the alert or that its start is not known. A project report names only the class of the failure. A note of a hypothesis gives the title of the finding and the title of the cause that comes first in the order of the hypotheses.

        Maximum length: 512

      • status.platformNotes.report
        string
        Name of the cluster report of the firing alert of the component; never in a project report.

        Maximum length: 253

      • status.platformNotes.since
        string
        The earliest moment the source proves the component not ready; absent when none is proven.
    • status.rejectedRerunNonce
      string
      Last rejected value of the annotation observability.deckhouse.io/rerun.

      Maximum length: 256

    • status.run
      object
      Current run.
      • status.run.attempt
        integer

        Required value

        Execution of the run; a new leader repeats an unfinished run with the next attempt, at most 3.

        Allowed values: 0 <= X

      • status.run.id
        string

        Required value

        Run ID.

        Maximum length: 64

      • status.run.linked
        array of objects
        Objects the run found from its targets, in the order they were found. The ids of the checks and built-in findings of such an object start with its checkPrefix; their outcomes do not change the phase of the run.
        • status.run.linked.apiVersion
          string
          API version.

          Maximum length: 128

        • status.run.linked.checkPrefix
          string
          Prefix of the ids of its checks and built-in findings: baseline/<n>-<kind>-<name>/, n is the order of the object, the name is cut to 40 characters.

          Maximum length: 320

        • status.run.linked.class
          string
          Platform for an object of Deckhouse: the label heritage: deckhouse on the object or on its namespace, a namespace d8-* or kube-system, a Node, or a cluster resource of an API group of Deckhouse; User for any other object. Written for a linked object that was read. Absent in a report written before the field existed.

          Allowed values: Platform, User

        • status.run.linked.kind
          string
          Kind.

          Maximum length: 63

        • status.run.linked.module
          string
          Module of Deckhouse of an object of class Platform, the label module of the object. Absent for class User and when neither names a module.

          Maximum length: 64

        • status.run.linked.moduleVersion
          string
          properties.version of the Module named by module. Absent when the Module was not found or could not be read, when the scope of the report does not read the Module (a project report whose class Platform rests on the labels of the object alone), and when the value is longer than 64 characters.

          Maximum length: 64

        • status.run.linked.name
          string
          Name.

          Maximum length: 253

        • status.run.linked.namespace
          string
          Namespace.

          Maximum length: 63

        • status.run.linked.of
          string
          checkPrefix of the object it was found from: baseline/ or baseline/<kind>-<name>/ of a target (baseline/<kind>-<namespace>-<name>/ when another target has its kind and name, with -<hash> at the end when two targets would share a prefix), or of another linked object.

          Maximum length: 320

        • status.run.linked.parentRef
          object
          Object it was found from, the target (an entry of status.run.targets) or the linked object (an entry of status.run.linked) whose prefix is of. Absent in a report written before the field existed.
          • status.run.linked.parentRef.apiVersion
            string
            API version.

            Maximum length: 128

          • status.run.linked.parentRef.kind
            string

            Required value

            Kind.

            Maximum length: 63

          • status.run.linked.parentRef.name
            string

            Required value

            Name.

            Maximum length: 253

          • status.run.linked.parentRef.namespace
            string
            Namespace.

            Maximum length: 63

          • status.run.linked.parentRef.resourceVersion
            string
            Resource version at the time of reading.

            Maximum length: 64

          • status.run.linked.parentRef.uid
            string
            UID.

            Maximum length: 64

        • status.run.linked.relation
          string
          Backend: the workload or the pod behind the Service an Ingress target routes to. Dependency: the object behind an address a log line connects to.

          Allowed values: Backend, Dependency

        • status.run.linked.stateCheckId
          string
          ID in status.checks of the row that holds the state of the object: <checkPrefix>target, written with the rows of the linked phase. The row is there also when the object was not found (its outcome tells). Absent when the run wrote no such row (the time or the checks of the run ran out before the object was read) and in a report written before the field existed.

          Maximum length: 320

        • status.run.linked.stoppedBy
          string
          Why the branch does not go on from the object; empty when it goes on.

          Allowed values: NoAddress, External, OutOfScope, Loop, MaxDepth, MaxObjects, TimeSlice, NotFound, MaxChecks, NotRead

        • status.run.linked.via
          object
          What the object was found through.
          • status.run.linked.via.check
            string

            Required value

            ID of the check of the object before it that named it.

            Maximum length: 320

          • status.run.linked.via.declaredIn
            string
            Where the configuration names the address: <container>: env <NAME> or <container>: args; the value is never stored.

            Maximum length: 384

          • status.run.linked.via.host
            string
            Host of the address of a log line, for a Dependency.

            Maximum length: 253

          • status.run.linked.via.port
            string
            Port of the address of a log line or of the configuration, for a Dependency.

            Maximum length: 5

          • status.run.linked.via.service
            string
            Service whose pods the object runs.

            Maximum length: 63

          • status.run.linked.via.source
            string
            Webhook for a Backend found through the configuration of a webhook; Config for a Dependency found by the configuration of a container (the literal values of its env, its command and args); omitted for one found by a log line.

            Allowed values: Config, Webhook

      • status.run.owner
        string
        Controller pod that executes the run.

        Maximum length: 253

      • status.run.playbooks
        array of objects
        Playbooks that took part in the run.
        • status.run.playbooks.fallback
          boolean
          true for a built-in playbook that ran on the object the labels of the alert name because no playbook names the alert; its findings describe that object, and a finding that would name a cause has the role ObjectProblem. Absent for a playbook of the alert and in a report written before the field existed.
        • status.run.playbooks.generation
          integer
          Generation.

          Allowed values: 0 <= X

        • status.run.playbooks.name
          string
          Name.

          Maximum length: 253

        • status.run.playbooks.source
          string
          Where the playbook comes from - Builtin, the image of the controller with generation 0, or Cluster, a ClusterObservabilityDiagnosticPlaybook.

          Allowed values: Builtin, Cluster

      • status.run.rerunNonce
        string
        Accepted value of the annotation observability.deckhouse.io/rerun.

        Maximum length: 256

      • status.run.startedAt
        string

        Required value

        Start of the run.
      • status.run.targetSource
        string
        Where the targets come from - Playbook, the target a playbook of the alert addresses, or AlertLabels, the object the labels of an alert that no playbook matches name, also when a playbook runs on it as a fallback. Empty for a run without a target.

        Allowed values: Playbook, AlertLabels

      • status.run.targets
        array of objects
        Diagnosed objects.
        • status.run.targets.apiVersion
          string
          API version.

          Maximum length: 128

        • status.run.targets.class
          string
          Platform for a Deckhouse object (label heritage: deckhouse on the object or its namespace, a d8-* or kube-system namespace, a Node, a cluster resource of a Deckhouse API group); User otherwise. When the object cannot be read, its namespace decides. Absent in older reports.

          Allowed values: Platform, User

        • status.run.targets.kind
          string
          Kind.

          Maximum length: 63

        • status.run.targets.module
          string
          Module of Deckhouse of an object of class Platform, the module of the component of its playbook, else the label module of the object, sanitized as the evidence is. Absent for class User, when neither names a module, and when the value is longer than 64 characters.

          Maximum length: 64

        • status.run.targets.moduleVersion
          string
          properties.version of the Module named by module. Absent when the Module was not found or could not be read, when the scope of the report does not read the Module (a project report whose class Platform rests on the labels of the object alone), and when the value is longer than 64 characters.

          Maximum length: 64

        • status.run.targets.name
          string
          Name.

          Maximum length: 253

        • status.run.targets.namespace
          string
          Namespace.

          Maximum length: 63

        • status.run.targets.resourceVersion
          string
          Resource version at the time of reading.

          Maximum length: 64

        • status.run.targets.stateCheckId
          string
          ID in status.checks of the row with the state of the object: baseline/target, or baseline/<kind>-<name>/target with several targets; present also when the object was not found (see its outcome). Absent until the row is written, when every playbook sets baseline: None, and in older reports.

          Maximum length: 320

        • status.run.targets.uid
          string
          UID.

          Maximum length: 64

      • status.run.trigger
        string

        Required value

        Alert for an automatic run, Manual for a report created by a user or a rerun.

        Allowed values: Alert, Manual

    • status.summary
      object
      Block for the UI.
      • status.summary.busy
        boolean
        A run is in progress.
      • status.summary.causeEstablished
        boolean
        true only when the headline is a finding that names a cause: not a symptom, not a fact and not a problem of the object (ObjectProblem), and Warning or Critical or of the role ConfigurationCause.
      • status.summary.causeGap
        string
        With causeEstablished false, one sentence in English on what the report could not read or follow.

        Maximum length: 300

      • status.summary.currentActivity
        string
        What the controller does or did, in English.

        Maximum length: 200

      • status.summary.findingsCount
        integer
        Number of findings.

        Allowed values: 0 <= X

      • status.summary.headline
        string
        Title of the main finding: the most severe one without causedBy that names a cause, else the most severe one without causedBy, else the most severe one; never a built-in fact. Starts with Cause not established: when it names no cause, e.g. Cause not established: no playbook matches the alert; with a fallback playbook, No playbook matches the alert: and the title of an ObjectProblem. Empty while the report is collecting.

        Maximum length: 300

      • status.summary.headlineFindingId
        string
        ID of the finding the headline comes from. Empty when no finding gives the headline.

        Maximum length: 320

      • status.summary.headlineSource
        string
        Finding when a finding gives the headline; empty otherwise.

        Allowed values: Finding

      • status.summary.progress
        object
        Progress of the collection.
        • status.summary.progress.checksDone
          integer
          Finished checks.

          Allowed values: 0 <= X

        • status.summary.progress.checksTotal
          integer
          Planned checks.

          Allowed values: 0 <= X

      • status.summary.severity
        string
        Highest severity of the findings.

        Allowed values: Info, Warning, Critical

      • status.summary.state
        string
        Collecting while the run goes, Ready when it is final, Failed when facts could not be collected. Derived from collection.phase.

        Allowed values: Collecting, Ready, Failed

      • status.summary.updatedAt
        string
        Time of the last update.

ClusterObservabilityLogsRulesGroup

Short names: colrg

Scope: Cluster
Version: v1alpha1

Group of log-based recording rules of the cluster scope.

Rules are evaluated by the logs-ruler service, which makes LogQL queries to the log storage backend. The resulting series are exposed as regular Prometheus metrics.

To set up an alert based on the created recording rules, use the ClusterObservabilityMetricsRulesGroup resource.

Used for monitoring logs from kube-*, d8-*, and cluster-level components.

  • spec
    object
    Desired state of a group of the log-based recording rules of the cluster scope.
    • spec.interval
      string

      Rule evaluation interval.

      If not specified, the default value is used.

      Default: 1m

      Pattern: ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$

    • spec.rules
      array of objects

      Required value

      List of log-based recording rules of the cluster scope.
      • spec.rules.expr
        string

        LogQL expression to evaluate the recording rule.

        The expression must explicitly select the intended log streams (for example, by namespace, app, or other labels available in the query backend).

        The namespace value is not injected automatically.

        Minimum length: 1

      • spec.rules.labels
        object

        Additional labels attached to the recorded metric series.

        The following labels are reserved and can’t be set by the user:

        • scope_type
        • created_by
        • rule_group_name
      • spec.rules.record
        string

        Name of the recording rule.

        Must comply with the Prometheus metric naming rules and be unique within the rule group.

        Minimum length: 1

        Pattern: ^[a-zA-Z_:][a-zA-Z0-9_:]*$

ClusterObservabilityMetricsRulesGroup

Short names: comrg

Scope: Cluster
Version: v1alpha1

Group of cluster-wide alerting or recording rules for system namespaces.

This resource lets platform administrators manage monitoring rules and alerts that apply to infrastructure and system namespaces (kube-*, d8-*, and an empty namespace).

  • spec
    object
    Defines the desired state of the cluster rules group.
    • spec.interval
      string

      Group rule evaluation interval.

      For example, 30s.

      If not specified, the default interval is used.

      Pattern: ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$

    • spec.rules
      array of objects

      Required value

      List of cluster-wide alerting or recording rules.
      • spec.rules.alert
        string

        Name of the alerting rule.

        Only one of the fields can be specified at the same time: alert or record.

        Pattern: ^[a-zA-Z-:][a-zA-Z0-9-:]*$

      • spec.rules.annotations
        object

        Annotations to attach to the alert.

        Applies only to alerting rules.

      • spec.rules.expr
        integer or string

        PromQL expression to evaluate the rule.

        Automatically scoped to infrastructure and system namespaces (kube-*, d8-*, and an empty namespace).

      • spec.rules.for
        string

        Delay before the alert is triggered.

        Defines how long the condition must continuously be true before the alert transitions to the firing state.

        Applies only to alerting rules.

        Pattern: ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$

      • spec.rules.keep_firing_for
        string

        Delay before the alert is resolved.

        Defines how long the alert remains in the firing state after the condition becomes false.

        Applies only to alerting rules.

        Pattern: ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$

      • spec.rules.labels
        object
        Additional labels to attach to the alert or recording rule.
      • spec.rules.record
        string

        Name of the recording rule.

        Only one of the fields can be specified at the same time: alert or record.

        Pattern: ^[a-zA-Z_:][a-zA-Z0-9_:]*$

  • status
    object
    State of the cluster rules group.
    • status.conditions
      array of objects
      The latest available observed state of the cluster rules group.
      • status.conditions.lastTransitionTime
        string
        Last time the condition transitioned from one status to another.
      • status.conditions.message
        string
        Human-readable message indicating details about last transition.
      • status.conditions.reason
        string

        Reason for the condition’s last transition.

        Specified in one word in CamelCase format.

      • status.conditions.status
        string

        Current status of the condition:

        • True: Condition is satisfied.
        • False: Condition is not satisfied.
        • Unknown: Condition could not be determined.

        Allowed values: True, False, Unknown

      • status.conditions.type
        string

        Type of the rules group condition:

        • Created: Resource is created.
        • Failed: Syncing failed.
        • Ready: Rules group is applied and active.

        Allowed values: Created, Failed, Ready

    • status.lastSyncTime
      string
      Last time a sync of the cluster rules group changed its PrometheusRule or its status. Not refreshed when a sync changes nothing.
    • status.prometheusRuleName
      string
      Name of the PrometheusRule resource created for this cluster rules group.

ClusterObservabilityNotificationChannel

Scope: Cluster
Version: v1alpha1

Notification channel for cluster-level alerts.

The module sends a notification when an alert fires and when it is resolved. Notification rules are set by a ClusterObservabilityNotificationPolicy object.

  • spec
    object

    Channel parameters.

    Set type and the section with the parameters of this type, for example, type: Email and email.

    • spec.email
      object
      Parameters of the Email channel.
      • spec.email.from
        string

        Required value

        Sender email address.
      • spec.email.smtp
        object

        Required value

        Parameters of the SMTP server that sends the emails.
        • spec.email.smtp.address
          string

          Required value

          Address of the SMTP server in the host:port format, for example, smtp.example.com:587.
        • spec.email.smtp.auth
          object
          Credentials for authentication on the SMTP server.
          • spec.email.smtp.auth.password
            string

            Required value

            Password.
          • spec.email.smtp.auth.username
            string

            Required value

            Username.
        • spec.email.smtp.requireTLS
          boolean
          Whether TLS verification is required.
      • spec.email.template
        string
        Email template.
    • spec.expressMessenger
      object
      Parameters of the ExpressMessenger channel.
      • spec.expressMessenger.apiUrl
        string

        Required value

        URL of the eXpress CTS server, for example, https://cts.company.ru.
      • spec.expressMessenger.botId
        string

        Required value

        ID (UUID) of the eXpress bot that sends the notifications.

        Pattern: ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$

      • spec.expressMessenger.secretKey
        string

        Required value

        Secret key of the bot for authentication (JWT v2).
      • spec.expressMessenger.template
        string
        Message template.
    • spec.slack
      object
      Parameters of the Slack channel.
      • spec.slack.apiUrl
        string

        Required value

        Slack API URL.
      • spec.slack.color
        string
        Color of the message attachment.
      • spec.slack.fallback
        string
        Fallback text of the message.
      • string
        Footer of the message.
      • spec.slack.iconEmoji
        string
        Emoji for the message.
      • spec.slack.iconUrl
        string
        URL of the icon image in the message.
      • spec.slack.pretext
        string
        Text above the main block of the message.
      • spec.slack.template
        string
        Message template.
      • spec.slack.text
        string
        Main text of the message.
      • spec.slack.title
        string
        Title of the message.
      • string
        URL that the title links to.
      • spec.slack.username
        string
        Name of the message sender.
    • spec.telegram
      object
      Parameters of the Telegram channel.
      • spec.telegram.apiToken
        string

        Required value

        API token of the Telegram bot that sends the notifications.
      • spec.telegram.apiUrl
        string
        Telegram Bot API URL.

        Default: https://api.telegram.org

      • spec.telegram.httpConfig
        object

        HTTP client parameters for requests to the Telegram API.

        Use proxyUrl or proxyFromEnvironment if the Telegram API is reachable only through a proxy.

        Set only one of the basicAuth, authorization and oauth2 parameters.

        • spec.telegram.httpConfig.authorization
          object
          Value of the Authorization header.
          • spec.telegram.httpConfig.authorization.credentials
            string
            Credentials inserted into the Authorization header.
          • spec.telegram.httpConfig.authorization.credentialsFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use credentials.

          • spec.telegram.httpConfig.authorization.type
            string

            Authentication scheme, for example, Bearer.

            If the parameter is omitted, Bearer is used.

        • spec.telegram.httpConfig.basicAuth
          object
          HTTP Basic authentication.
          • spec.telegram.httpConfig.basicAuth.password
            string
            Password.
          • spec.telegram.httpConfig.basicAuth.passwordFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use password.

          • spec.telegram.httpConfig.basicAuth.username
            string
            Username.
        • spec.telegram.httpConfig.enableHttp2
          boolean
          Enable HTTP/2 support.
        • spec.telegram.httpConfig.followRedirects
          boolean

          Deprecated and ignored for security reasons.

          The module does not follow redirects because a redirect makes Alertmanager send a request to an address chosen by the notification recipient.

        • spec.telegram.httpConfig.httpHeaders
          object

          Custom HTTP headers of requests to the Telegram API.

          The object key is the header name, the value is an object with a list of values and/or references to secrets and files.

          • spec.telegram.httpConfig.httpHeaders.<KEY_NAME>
            object
            <KEY_NAME> — item (key) name.
            • spec.telegram.httpConfig.httpHeaders.<KEY_NAME>.files
              array of strings

              Deprecated and ignored for security reasons.

              File contents are not inserted into request headers. Use values or secrets.

            • spec.telegram.httpConfig.httpHeaders.<KEY_NAME>.secrets
              array of strings
              Header values read from secrets.
            • spec.telegram.httpConfig.httpHeaders.<KEY_NAME>.values
              array of strings
              Header values.
        • spec.telegram.httpConfig.noProxy
          string
          Comma-separated list of hosts for which the proxy is not used.
        • spec.telegram.httpConfig.oauth2
          object
          OAuth 2.0 authentication.
          • spec.telegram.httpConfig.oauth2.clientId
            string

            Required value

            OAuth 2.0 client ID.
          • spec.telegram.httpConfig.oauth2.clientSecret
            string
            OAuth 2.0 client secret.
          • spec.telegram.httpConfig.oauth2.clientSecretFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use clientSecret.

          • spec.telegram.httpConfig.oauth2.endpointParams
            object
            Additional parameters passed in the token request.
          • spec.telegram.httpConfig.oauth2.noProxy
            string
            Comma-separated list of hosts for which the proxy is not used.
          • spec.telegram.httpConfig.oauth2.proxyFromEnvironment
            boolean
            Use the proxy from the HTTP_PROXY, HTTPS_PROXY and NO_PROXY environment variables.
          • spec.telegram.httpConfig.oauth2.proxyUrl
            string
            URL of the HTTP proxy used for the token request.
          • spec.telegram.httpConfig.oauth2.scopes
            array of strings
            Scopes requested for the token.
            • spec.telegram.httpConfig.oauth2.scopes.Element of the array
              string
              Scope name.
          • spec.telegram.httpConfig.oauth2.tlsConfig
            object
            TLS settings for the token request.
            • spec.telegram.httpConfig.oauth2.tlsConfig.caFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.telegram.httpConfig.oauth2.tlsConfig.certFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.telegram.httpConfig.oauth2.tlsConfig.insecureSkipVerify
              boolean
              Disable verification of the server TLS certificate.
            • spec.telegram.httpConfig.oauth2.tlsConfig.keyFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.telegram.httpConfig.oauth2.tlsConfig.maxVersion
              string
              Maximum TLS version.
            • spec.telegram.httpConfig.oauth2.tlsConfig.minVersion
              string
              Minimum TLS version.
            • spec.telegram.httpConfig.oauth2.tlsConfig.serverName
              string
              Server name for TLS certificate verification (SNI).
          • spec.telegram.httpConfig.oauth2.tokenUrl
            string

            Required value

            URL of the token endpoint.
        • spec.telegram.httpConfig.proxyConnectHeader
          object

          Headers of CONNECT requests to the proxy.

          Header names are mapped to lists of values.

        • spec.telegram.httpConfig.proxyFromEnvironment
          boolean
          Use the proxy from the HTTP_PROXY, HTTPS_PROXY and NO_PROXY environment variables.
        • spec.telegram.httpConfig.proxyUrl
          string
          URL of the HTTP proxy for requests to the Telegram API.
        • spec.telegram.httpConfig.tlsConfig
          object
          TLS settings for the HTTP client of the Telegram receiver.
          • spec.telegram.httpConfig.tlsConfig.caFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.telegram.httpConfig.tlsConfig.certFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.telegram.httpConfig.tlsConfig.insecureSkipVerify
            boolean
            Disable verification of the server TLS certificate.
          • spec.telegram.httpConfig.tlsConfig.keyFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.telegram.httpConfig.tlsConfig.maxVersion
            string
            Maximum TLS version.
          • spec.telegram.httpConfig.tlsConfig.minVersion
            string
            Minimum TLS version.
          • spec.telegram.httpConfig.tlsConfig.serverName
            string
            Server name for TLS certificate verification (SNI).
      • spec.telegram.template
        string
        Message template.
    • spec.type
      string

      Required value

      Notification channel type.

      Allowed values: Email, Slack, Telegram, Webhook, ExpressMessenger, Zabbix

    • spec.webhook
      object
      Parameters of the Webhook channel.
      • spec.webhook.httpConfig
        object

        HTTP client parameters for requests to the webhook URL.

        Set only one of the basicAuth, authorization and oauth2 parameters.

        • spec.webhook.httpConfig.authorization
          object
          Value of the Authorization header.
          • spec.webhook.httpConfig.authorization.credentials
            string
            Credentials inserted into the Authorization header.
          • spec.webhook.httpConfig.authorization.credentialsFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use credentials.

          • spec.webhook.httpConfig.authorization.type
            string

            Authentication scheme, for example, Bearer.

            If the parameter is omitted, Bearer is used.

        • spec.webhook.httpConfig.basicAuth
          object
          HTTP Basic authentication.
          • spec.webhook.httpConfig.basicAuth.password
            string
            Password.
          • spec.webhook.httpConfig.basicAuth.passwordFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use password.

          • spec.webhook.httpConfig.basicAuth.username
            string
            Username.
        • spec.webhook.httpConfig.enableHttp2
          boolean
          Enable HTTP/2 support for webhook requests.
        • spec.webhook.httpConfig.followRedirects
          boolean

          Deprecated and ignored for security reasons.

          The module does not follow HTTP redirects.

        • spec.webhook.httpConfig.httpHeaders
          object

          Custom HTTP headers of requests to the webhook URL.

          The object key is the header name, the value is an object with a list of values and/or references to secrets and files.

          • spec.webhook.httpConfig.httpHeaders.<KEY_NAME>
            object
            <KEY_NAME> — item (key) name.
            • spec.webhook.httpConfig.httpHeaders.<KEY_NAME>.files
              array of strings

              Deprecated and ignored for security reasons.

              File contents are not inserted into request headers. Use values or secrets.

            • spec.webhook.httpConfig.httpHeaders.<KEY_NAME>.secrets
              array of strings
              Header values read from secrets.
            • spec.webhook.httpConfig.httpHeaders.<KEY_NAME>.values
              array of strings
              Header values.
        • spec.webhook.httpConfig.noProxy
          string
          Comma-separated list of hosts that are accessed without the proxy.
        • spec.webhook.httpConfig.oauth2
          object
          OAuth 2.0 authentication.
          • spec.webhook.httpConfig.oauth2.clientId
            string

            Required value

            OAuth 2.0 client ID.
          • spec.webhook.httpConfig.oauth2.clientSecret
            string
            OAuth 2.0 client secret.
          • spec.webhook.httpConfig.oauth2.clientSecretFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use clientSecret.

          • spec.webhook.httpConfig.oauth2.endpointParams
            object
            Additional parameters passed in the token request.
          • spec.webhook.httpConfig.oauth2.noProxy
            string
            Comma-separated list of hosts for which the proxy is not used.
          • spec.webhook.httpConfig.oauth2.proxyFromEnvironment
            boolean
            Use the proxy from the HTTP_PROXY, HTTPS_PROXY and NO_PROXY environment variables.
          • spec.webhook.httpConfig.oauth2.proxyUrl
            string
            URL of the HTTP proxy used for the token request.
          • spec.webhook.httpConfig.oauth2.scopes
            array of strings
            Scopes requested for the token.
            • spec.webhook.httpConfig.oauth2.scopes.Element of the array
              string
              Scope name.
          • spec.webhook.httpConfig.oauth2.tlsConfig
            object
            TLS settings for the token request.
            • spec.webhook.httpConfig.oauth2.tlsConfig.caFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.webhook.httpConfig.oauth2.tlsConfig.certFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.webhook.httpConfig.oauth2.tlsConfig.insecureSkipVerify
              boolean
              Disable verification of the server TLS certificate.
            • spec.webhook.httpConfig.oauth2.tlsConfig.keyFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.webhook.httpConfig.oauth2.tlsConfig.maxVersion
              string
              Maximum TLS version.
            • spec.webhook.httpConfig.oauth2.tlsConfig.minVersion
              string
              Minimum TLS version.
            • spec.webhook.httpConfig.oauth2.tlsConfig.serverName
              string
              Server name for TLS certificate verification (SNI).
          • spec.webhook.httpConfig.oauth2.tokenUrl
            string

            Required value

            URL of the token endpoint.
        • spec.webhook.httpConfig.proxyConnectHeader
          object

          Headers of CONNECT requests to the proxy.

          Header names are mapped to lists of values.

        • spec.webhook.httpConfig.proxyFromEnvironment
          boolean
          Use the proxy parameters from the HTTP_PROXY, HTTPS_PROXY and NO_PROXY environment variables.
        • spec.webhook.httpConfig.proxyUrl
          string
          URL of the HTTP proxy for webhook requests.
        • spec.webhook.httpConfig.tlsConfig
          object
          TLS settings for requests to the webhook URL.
          • spec.webhook.httpConfig.tlsConfig.caFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.webhook.httpConfig.tlsConfig.certFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.webhook.httpConfig.tlsConfig.insecureSkipVerify
            boolean
            Disable verification of the server TLS certificate.
          • spec.webhook.httpConfig.tlsConfig.keyFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.webhook.httpConfig.tlsConfig.maxVersion
            string
            Maximum TLS version.
          • spec.webhook.httpConfig.tlsConfig.minVersion
            string
            Minimum TLS version.
          • spec.webhook.httpConfig.tlsConfig.serverName
            string
            Server name for TLS certificate verification (SNI).
      • spec.webhook.url
        string

        Required value

        Webhook URL.

        The module sends notifications to this URL as JSON in the Alertmanager webhook format.

        Minimum length: 1

    • spec.zabbix
      object
      Parameters of the Zabbix channel.
      • spec.zabbix.api
        object

        Connection to the Zabbix API for automatic import of the template.

        If the section is set, the module imports the template and creates the cluster host and host group through the Zabbix JSON-RPC API. Otherwise, import the template from the ConfigMap manually.

        • spec.zabbix.api.tlsConfig
          object

          TLS settings of the HTTPS connection to the Zabbix API.

          Certificates and keys are given inline as PEM.

          • spec.zabbix.api.tlsConfig.ca
            string
            PEM-encoded CA certificate(s) used to verify the Zabbix API server certificate. If unset, the system trust store is used.
          • spec.zabbix.api.tlsConfig.caFile
            Deprecated
            string
            Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into ca instead.
          • spec.zabbix.api.tlsConfig.cert
            string
            PEM-encoded client certificate, for an API behind mutual TLS. Must be set together with key.
          • spec.zabbix.api.tlsConfig.certFile
            Deprecated
            string
            Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into cert instead.
          • spec.zabbix.api.tlsConfig.insecureSkipVerify
            boolean
            Disable verification of the Zabbix API server TLS certificate.
          • spec.zabbix.api.tlsConfig.key
            string
            PEM-encoded unencrypted private key for cert. Must be set together with cert.
          • spec.zabbix.api.tlsConfig.keyFile
            Deprecated
            string
            Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into key instead.
          • spec.zabbix.api.tlsConfig.serverName
            string
            Server name used for TLS certificate verification (SNI). Defaults to the host of url.
        • spec.zabbix.api.token
          string

          Required value

          Zabbix API token used to authenticate configuration.import calls.

          Minimum length: 1

        • spec.zabbix.api.url
          string

          Required value

          URL of the Zabbix JSON-RPC API, for example, https://zabbix.example.com/api_jsonrpc.php.

          If the scheme is omitted, https:// is used.

          Minimum length: 1

      • spec.zabbix.clusterName
        string

        Required value

        Name of the cluster in Zabbix.

        The name becomes the technical name of the Zabbix host that receives the data. It must be unique among all clusters and channels that send data to the same Zabbix instance.

        Minimum length: 1

      • spec.zabbix.disableTLS
        boolean

        Connect to the trapper port without TLS.

        If the parameter is disabled, the connection uses TLS with the settings from tls. If tls is omitted, the certificate is verified.

        Default: false

      • spec.zabbix.hostGroup
        string
        Zabbix host group that the cluster host is added to.

        Default: Deckhouse

      • spec.zabbix.port
        integer
        Trapper port of the Zabbix Server or Zabbix Proxy.

        Default: 10051

      • spec.zabbix.server
        string

        Required value

        Address of the Zabbix Server or Zabbix Proxy that receives trapper data.

        Minimum length: 1

      • spec.zabbix.tls
        object

        TLS settings of the connection to the Zabbix Server over the sender protocol.

        Certificates and keys are given inline as PEM. Zabbix certificate mode (TLSAccept=cert) is mutual TLS: cert and key are mandatory for it. Zabbix PSK mode is not supported.

        • spec.zabbix.tls.ca
          string
          PEM-encoded CA certificate(s) used to verify the Zabbix Server/Proxy certificate. If unset, the system trust store is used.
        • spec.zabbix.tls.caFile
          Deprecated
          string
          Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into ca instead.
        • spec.zabbix.tls.cert
          string
          PEM-encoded client certificate presented to the Zabbix Server/Proxy. Must be set together with key.
        • spec.zabbix.tls.certFile
          Deprecated
          string
          Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into cert instead.
        • spec.zabbix.tls.insecureSkipVerify
          boolean
          Disable verification of the Zabbix server TLS certificate entirely. Does not lift the client certificate requirement of Zabbix certificate mode.
        • spec.zabbix.tls.key
          string
          PEM-encoded unencrypted private key for cert. Must be set together with cert.
        • spec.zabbix.tls.keyFile
          Deprecated
          string
          Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into key instead.
        • spec.zabbix.tls.serverName
          string
          Server name used for TLS certificate verification (SNI). Defaults to spec.zabbix.server.
        • spec.zabbix.tls.skipServerNameVerification
          boolean

          Verify the server certificate chain against ca but do not match the certificate against serverName.

          Needed when the Zabbix certificate has no Subject Alternative Name (Zabbix validates only the Subject, so this is common), which is otherwise rejected. Requires ca: against the system trust store alone, a chain check without a name would accept any publicly trusted certificate. Ignored when insecureSkipVerify is true.

  • status
    object
    Status of the notification channel, including which policies are using it.
    • status.policies
      array of objects
      List of notification policies that are using this channel across all namespaces.
      • status.policies.name
        string
        Name of the notification policy.
      • status.policies.namespace
        string

        Namespace of the notification policy.

        Empty string for ClusterObservabilityNotificationPolicy.

ClusterObservabilityNotificationPolicy

Scope: Cluster
Version: v1alpha1

Rules for sending notifications about cluster-level alerts.

Using the policy, the module selects cluster-level alerts by their labels and sends notifications about them to a ClusterObservabilityNotificationChannel.

  • spec
    object
    Policy parameters.
    • spec.alert
      object

      Required value

      Alerts that the policy applies to.
      • spec.alert.groupByLabels
        array of strings

        Labels by which alerts are grouped into one notification.

        Alerts with the same values of these labels are sent in one message. If the parameter is omitted, a separate notification is sent for each alert.

        • spec.alert.groupByLabels.Element of the array
          string
          Alert label name.
      • spec.alert.selector
        object

        Required value

        Selector of the alerts that the policy applies to.

        Set only one of the fields: matchLabels or matchExpressions.

        • spec.alert.selector.matchExpressions
          array of objects

          Set of expressions (conditions) for filtering alerts.

          The policy applies to an alert if the alert meets all of the conditions.

          • spec.alert.selector.matchExpressions.key
            string
            Alert label name.
          • spec.alert.selector.matchExpressions.operator
            string

            Comparison operator:

            • Equal: The label value equals value.
            • NotEqual: The label value differs from value.
            • Regex: The label value matches the value regular expression.
            • NotRegex: The label value does not match the value regular expression.

            Allowed values: Equal, NotEqual, Regex, NotRegex

          • spec.alert.selector.matchExpressions.value
            string
            Value or regular expression to compare the label value with.
        • spec.alert.selector.matchLabels
          object
          Labels and their values that an alert must have for the policy to apply to it.
    • spec.notification
      object

      Required value

      Notification parameters.
      • spec.notification.channel
        object

        Required value

        “Channel that receives the notifications.”
        • spec.notification.channel.kind
          string

          Required value

          Kind of the notification channel object.

          Default: ClusterObservabilityNotificationChannel

          Allowed values: ClusterObservabilityNotificationChannel

        • spec.notification.channel.name
          string

          Required value

          Name of the channel object.
        • spec.notification.channel.parameters
          object

          Required value

          Notification channel parameters.
      • spec.notification.repeatInterval
        string

        Required value

        Interval for resending the notification about an alert.

        Default: 1m

  • status
    object
    Notification policy state.
    • status.description
      string
      Reason why the module does not apply the policy.
    • status.state
      string

      Policy state:

      • Applicable: The module sends notifications according to the policy.
      • NotApplicable: The module does not apply the policy, for example, because the channel does not exist or parameters do not match the channel type. The reason is in the description field.

ClusterObservabilityNotificationSilence

Scope: Cluster
Version: v1alpha1

Temporary silence of notifications about cluster-level alerts.

During the silence, the module does not send notifications about cluster-level alerts if they match spec.selector.

  • spec
    object
    Silence parameters.
    • spec.endsAt
      string

      Required value

      Time when the silence ends.
    • spec.selector
      object

      Required value

      Selector of the alerts whose notifications are silenced.

      Set only one of the fields: matchLabels or matchExpressions.

      • spec.selector.matchExpressions
        array of objects

        Set of expressions (conditions) for filtering alerts.

        Notifications about an alert are silenced if the alert meets all of the conditions.

        • spec.selector.matchExpressions.key
          string
          Alert label name.
        • spec.selector.matchExpressions.operator
          string

          Comparison operator:

          • Equal: The label value equals value.
          • NotEqual: The label value differs from value.
          • Regex: The label value matches the value regular expression.
          • NotRegex: The label value does not match the value regular expression.

          Allowed values: Equal, NotEqual, Regex, NotRegex

        • spec.selector.matchExpressions.value
          string
          Value or regular expression to compare the label value with.
      • spec.selector.matchLabels
        object
        Labels and their values that an alert must have for its notifications to be silenced.
    • spec.startsAt
      string

      Required value

      Time when the silence starts.

ClusterObservabilityPropagatedDashboard

Scope: Cluster
Version: v1alpha1

Dashboard displayed in the Deckhouse web UI at the cluster level and in all namespaces (projects).

To configure how the dashboard is displayed, use the following annotations:

  • metadata.deckhouse.io/category: Folder (category) of the dashboard in the interface.
  • metadata.deckhouse.io/title: Title of the dashboard in the interface. If the annotation is omitted, the module uses the title from the JSON manifest.
  • spec
    object
    • spec.definition
      string

      Required value

      Dashboard JSON manifest as a string compatible with the Grafana format.

ClusterObservabilityPropagatedMetricsRulesGroup

Short names: copmrg

Scope: Cluster
Version: v1alpha1

Cluster-wide group of alerting and recording rules propagated to all namespaces.

Resources with the heritage: deckhouse label are shipped with Deckhouse Platform and can be managed only by the deckhouse ServiceAccount.

  • spec
    object
    Propagated rule group parameters.
    • spec.interval
      string

      Group rule evaluation interval.

      For example, 30s.

      If not specified, the default interval is used.

      Pattern: ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$

    • spec.rules
      array of objects

      Required value

      List of alerting and recording rules of the group.
      • spec.rules.alert
        string

        Name of the alerting rule.

        Only one of the fields can be specified at the same time: alert or record.

        Pattern: ^[a-zA-Z-:][a-zA-Z0-9-:]*$

      • spec.rules.annotations
        object

        Annotations to attach to the alert.

        Applies only to alerting rules.

      • spec.rules.expr
        integer or string

        PromQL expression to evaluate the rule.

        Propagated rules aren’t restricted by namespaces.

      • spec.rules.for
        string

        Delay before the alert is triggered.

        Defines how long the condition must continuously be true before the alert transitions to the firing state.

        Applies only to alerting rules.

        Pattern: ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$

      • spec.rules.keep_firing_for
        string

        Delay before the alert is resolved.

        Defines how long the alert remains in the firing state after the condition becomes false.

        Applies only to alerting rules.

        Pattern: ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$

      • spec.rules.labels
        object
        Additional labels to attach to the alert or recording rule.
      • spec.rules.record
        string

        Name of the recording rule.

        Only one of the fields can be specified at the same time: alert or record.

        Pattern: ^[a-zA-Z_:][a-zA-Z0-9_:]*$

  • status
    object
    State of the cluster rules group.
    • status.conditions
      array of objects
      The latest available observed state of the propagated rules group.
      • status.conditions.lastTransitionTime
        string
        Last time the condition transitioned from one status to another.
      • status.conditions.message
        string
        Human-readable message indicating details about last transition.
      • status.conditions.reason
        string

        Reason for the condition’s last transition.

        Specified in one word in CamelCase format.

      • status.conditions.status
        string

        Current status of the condition:

        • True: Condition is satisfied.
        • False: Condition is not satisfied.
        • Unknown: Condition could not be determined.

        Allowed values: True, False, Unknown

      • status.conditions.type
        string

        Type of the rules group condition:

        • Created: Resource is created.
        • Failed: Syncing failed.
        • Ready: Rules group is applied and active.

        Allowed values: Created, Failed, Ready

    • status.lastSyncTime
      string

      Last time the propagated rules group was successfully synced, which changed its PrometheusRule or its status.

      Not refreshed when a sync changes nothing.

    • status.prometheusRuleName
      string
      Name of the PrometheusRule resource created for this propagated rules group.

ClusterObservabilityPropagatedNotificationChannel

Scope: Cluster
Version: v1alpha1

Notification channel available in all namespaces.

The module sends a notification when an alert fires and when it is resolved. Notification rules are set by ObservabilityNotificationPolicy objects in any namespace or by a ClusterObservabilityPropagatedNotificationPolicy object.

  • spec
    object

    Channel parameters.

    Set type and the section with the parameters of this type, for example, type: Email and email.

    • spec.email
      object
      Parameters of the Email channel.
      • spec.email.from
        string

        Required value

        Sender email address.
      • spec.email.smtp
        object

        Required value

        Parameters of the SMTP server that sends the emails.
        • spec.email.smtp.address
          string

          Required value

          Address of the SMTP server in the host:port format, for example, smtp.example.com:587.
        • spec.email.smtp.auth
          object
          Credentials for authentication on the SMTP server.
          • spec.email.smtp.auth.password
            string

            Required value

            Password.
          • spec.email.smtp.auth.username
            string

            Required value

            Username.
        • spec.email.smtp.requireTLS
          boolean
          Whether TLS verification is required.
      • spec.email.template
        string
        Email template.
    • spec.expressMessenger
      object
      Parameters of the ExpressMessenger channel.
      • spec.expressMessenger.apiUrl
        string

        Required value

        URL of the eXpress CTS server, for example, https://cts.company.ru.
      • spec.expressMessenger.botId
        string

        Required value

        ID (UUID) of the eXpress bot that sends the notifications.

        Pattern: ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$

      • spec.expressMessenger.secretKey
        string

        Required value

        Secret key of the bot for authentication (JWT v2).
      • spec.expressMessenger.template
        string
        Message template.
    • spec.slack
      object
      Parameters of the Slack channel.
      • spec.slack.apiUrl
        string

        Required value

        Slack API URL.
      • spec.slack.color
        string
        Color of the message attachment.
      • spec.slack.fallback
        string
        Fallback text of the message.
      • string
        Footer of the message.
      • spec.slack.iconEmoji
        string
        Emoji for the message.
      • spec.slack.iconUrl
        string
        URL of the icon image in the message.
      • spec.slack.pretext
        string
        Text above the main block of the message.
      • spec.slack.template
        string
        Message template.
      • spec.slack.text
        string
        Main text of the message.
      • spec.slack.title
        string
        Title of the message.
      • string
        URL that the title links to.
      • spec.slack.username
        string
        Name of the message sender.
    • spec.telegram
      object
      Parameters of the Telegram channel.
      • spec.telegram.apiToken
        string

        Required value

        API token of the Telegram bot that sends the notifications.
      • spec.telegram.apiUrl
        string
        Telegram Bot API URL.

        Default: https://api.telegram.org

      • spec.telegram.httpConfig
        object

        HTTP client parameters for requests to the Telegram API.

        Use proxyUrl or proxyFromEnvironment if the Telegram API is reachable only through a proxy.

        Set only one of the basicAuth, authorization and oauth2 parameters.

        • spec.telegram.httpConfig.authorization
          object
          Value of the Authorization header.
          • spec.telegram.httpConfig.authorization.credentials
            string
            Credentials inserted into the Authorization header.
          • spec.telegram.httpConfig.authorization.credentialsFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use credentials.

          • spec.telegram.httpConfig.authorization.type
            string

            Authentication scheme, for example, Bearer.

            If the parameter is omitted, Bearer is used.

        • spec.telegram.httpConfig.basicAuth
          object
          HTTP Basic authentication.
          • spec.telegram.httpConfig.basicAuth.password
            string
            Password.
          • spec.telegram.httpConfig.basicAuth.passwordFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use password.

          • spec.telegram.httpConfig.basicAuth.username
            string
            Username.
        • spec.telegram.httpConfig.enableHttp2
          boolean
          Enable HTTP/2 support.
        • spec.telegram.httpConfig.followRedirects
          boolean

          Deprecated and ignored for security reasons.

          The module does not follow redirects because a redirect makes Alertmanager send a request to an address chosen by the notification recipient.

        • spec.telegram.httpConfig.httpHeaders
          object

          Custom HTTP headers of requests to the Telegram API.

          The object key is the header name, the value is an object with a list of values and/or references to secrets and files.

          • spec.telegram.httpConfig.httpHeaders.<KEY_NAME>
            object
            <KEY_NAME> — item (key) name.
            • spec.telegram.httpConfig.httpHeaders.<KEY_NAME>.files
              array of strings

              Deprecated and ignored for security reasons.

              File contents are not inserted into request headers. Use values or secrets.

            • spec.telegram.httpConfig.httpHeaders.<KEY_NAME>.secrets
              array of strings
              Header values read from secrets.
            • spec.telegram.httpConfig.httpHeaders.<KEY_NAME>.values
              array of strings
              Header values.
        • spec.telegram.httpConfig.noProxy
          string
          Comma-separated list of hosts for which the proxy is not used.
        • spec.telegram.httpConfig.oauth2
          object
          OAuth 2.0 authentication.
          • spec.telegram.httpConfig.oauth2.clientId
            string

            Required value

            OAuth 2.0 client ID.
          • spec.telegram.httpConfig.oauth2.clientSecret
            string
            OAuth 2.0 client secret.
          • spec.telegram.httpConfig.oauth2.clientSecretFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use clientSecret.

          • spec.telegram.httpConfig.oauth2.endpointParams
            object
            Additional parameters passed in the token request.
          • spec.telegram.httpConfig.oauth2.noProxy
            string
            Comma-separated list of hosts for which the proxy is not used.
          • spec.telegram.httpConfig.oauth2.proxyFromEnvironment
            boolean
            Use the proxy from the HTTP_PROXY, HTTPS_PROXY and NO_PROXY environment variables.
          • spec.telegram.httpConfig.oauth2.proxyUrl
            string
            URL of the HTTP proxy used for the token request.
          • spec.telegram.httpConfig.oauth2.scopes
            array of strings
            Scopes requested for the token.
            • spec.telegram.httpConfig.oauth2.scopes.Element of the array
              string
              Scope name.
          • spec.telegram.httpConfig.oauth2.tlsConfig
            object
            TLS settings for the token request.
            • spec.telegram.httpConfig.oauth2.tlsConfig.caFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.telegram.httpConfig.oauth2.tlsConfig.certFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.telegram.httpConfig.oauth2.tlsConfig.insecureSkipVerify
              boolean
              Disable verification of the server TLS certificate.
            • spec.telegram.httpConfig.oauth2.tlsConfig.keyFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.telegram.httpConfig.oauth2.tlsConfig.maxVersion
              string
              Maximum TLS version.
            • spec.telegram.httpConfig.oauth2.tlsConfig.minVersion
              string
              Minimum TLS version.
            • spec.telegram.httpConfig.oauth2.tlsConfig.serverName
              string
              Server name for TLS certificate verification (SNI).
          • spec.telegram.httpConfig.oauth2.tokenUrl
            string

            Required value

            URL of the token endpoint.
        • spec.telegram.httpConfig.proxyConnectHeader
          object

          Headers of CONNECT requests to the proxy.

          Header names are mapped to lists of values.

        • spec.telegram.httpConfig.proxyFromEnvironment
          boolean
          Use the proxy from the HTTP_PROXY, HTTPS_PROXY and NO_PROXY environment variables.
        • spec.telegram.httpConfig.proxyUrl
          string
          URL of the HTTP proxy for requests to the Telegram API.
        • spec.telegram.httpConfig.tlsConfig
          object
          TLS settings for the HTTP client of the Telegram receiver.
          • spec.telegram.httpConfig.tlsConfig.caFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.telegram.httpConfig.tlsConfig.certFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.telegram.httpConfig.tlsConfig.insecureSkipVerify
            boolean
            Disable verification of the server TLS certificate.
          • spec.telegram.httpConfig.tlsConfig.keyFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.telegram.httpConfig.tlsConfig.maxVersion
            string
            Maximum TLS version.
          • spec.telegram.httpConfig.tlsConfig.minVersion
            string
            Minimum TLS version.
          • spec.telegram.httpConfig.tlsConfig.serverName
            string
            Server name for TLS certificate verification (SNI).
      • spec.telegram.template
        string
        Message template.
    • spec.type
      string

      Required value

      Notification channel type.

      Allowed values: Email, Slack, Telegram, Webhook, ExpressMessenger, Zabbix

    • spec.webhook
      object
      Parameters of the Webhook channel.
      • spec.webhook.httpConfig
        object

        HTTP client parameters for requests to the webhook URL.

        Set only one of the basicAuth, authorization and oauth2 parameters.

        • spec.webhook.httpConfig.authorization
          object
          Value of the Authorization header.
          • spec.webhook.httpConfig.authorization.credentials
            string
            Credentials inserted into the Authorization header.
          • spec.webhook.httpConfig.authorization.credentialsFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use credentials.

          • spec.webhook.httpConfig.authorization.type
            string

            Authentication scheme, for example, Bearer.

            If the parameter is omitted, Bearer is used.

        • spec.webhook.httpConfig.basicAuth
          object
          HTTP Basic authentication.
          • spec.webhook.httpConfig.basicAuth.password
            string
            Password.
          • spec.webhook.httpConfig.basicAuth.passwordFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use password.

          • spec.webhook.httpConfig.basicAuth.username
            string
            Username.
        • spec.webhook.httpConfig.enableHttp2
          boolean
          Enable HTTP/2 support for webhook requests.
        • spec.webhook.httpConfig.followRedirects
          boolean

          Deprecated and ignored for security reasons.

          The module does not follow HTTP redirects.

        • spec.webhook.httpConfig.httpHeaders
          object

          Custom HTTP headers of requests to the webhook URL.

          The object key is the header name, the value is an object with a list of values and/or references to secrets and files.

          • spec.webhook.httpConfig.httpHeaders.<KEY_NAME>
            object
            <KEY_NAME> — item (key) name.
            • spec.webhook.httpConfig.httpHeaders.<KEY_NAME>.files
              array of strings

              Deprecated and ignored for security reasons.

              File contents are not inserted into request headers. Use values or secrets.

            • spec.webhook.httpConfig.httpHeaders.<KEY_NAME>.secrets
              array of strings
              Header values read from secrets.
            • spec.webhook.httpConfig.httpHeaders.<KEY_NAME>.values
              array of strings
              Header values.
        • spec.webhook.httpConfig.noProxy
          string
          Comma-separated list of hosts that are accessed without the proxy.
        • spec.webhook.httpConfig.oauth2
          object
          OAuth 2.0 authentication.
          • spec.webhook.httpConfig.oauth2.clientId
            string

            Required value

            OAuth 2.0 client ID.
          • spec.webhook.httpConfig.oauth2.clientSecret
            string
            OAuth 2.0 client secret.
          • spec.webhook.httpConfig.oauth2.clientSecretFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use clientSecret.

          • spec.webhook.httpConfig.oauth2.endpointParams
            object
            Additional parameters passed in the token request.
          • spec.webhook.httpConfig.oauth2.noProxy
            string
            Comma-separated list of hosts for which the proxy is not used.
          • spec.webhook.httpConfig.oauth2.proxyFromEnvironment
            boolean
            Use the proxy from the HTTP_PROXY, HTTPS_PROXY and NO_PROXY environment variables.
          • spec.webhook.httpConfig.oauth2.proxyUrl
            string
            URL of the HTTP proxy used for the token request.
          • spec.webhook.httpConfig.oauth2.scopes
            array of strings
            Scopes requested for the token.
            • spec.webhook.httpConfig.oauth2.scopes.Element of the array
              string
              Scope name.
          • spec.webhook.httpConfig.oauth2.tlsConfig
            object
            TLS settings for the token request.
            • spec.webhook.httpConfig.oauth2.tlsConfig.caFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.webhook.httpConfig.oauth2.tlsConfig.certFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.webhook.httpConfig.oauth2.tlsConfig.insecureSkipVerify
              boolean
              Disable verification of the server TLS certificate.
            • spec.webhook.httpConfig.oauth2.tlsConfig.keyFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.webhook.httpConfig.oauth2.tlsConfig.maxVersion
              string
              Maximum TLS version.
            • spec.webhook.httpConfig.oauth2.tlsConfig.minVersion
              string
              Minimum TLS version.
            • spec.webhook.httpConfig.oauth2.tlsConfig.serverName
              string
              Server name for TLS certificate verification (SNI).
          • spec.webhook.httpConfig.oauth2.tokenUrl
            string

            Required value

            URL of the token endpoint.
        • spec.webhook.httpConfig.proxyConnectHeader
          object

          Headers of CONNECT requests to the proxy.

          Header names are mapped to lists of values.

        • spec.webhook.httpConfig.proxyFromEnvironment
          boolean
          Use the proxy parameters from the HTTP_PROXY, HTTPS_PROXY and NO_PROXY environment variables.
        • spec.webhook.httpConfig.proxyUrl
          string
          URL of the HTTP proxy for webhook requests.
        • spec.webhook.httpConfig.tlsConfig
          object
          TLS settings for requests to the webhook URL.
          • spec.webhook.httpConfig.tlsConfig.caFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.webhook.httpConfig.tlsConfig.certFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.webhook.httpConfig.tlsConfig.insecureSkipVerify
            boolean
            Disable verification of the server TLS certificate.
          • spec.webhook.httpConfig.tlsConfig.keyFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.webhook.httpConfig.tlsConfig.maxVersion
            string
            Maximum TLS version.
          • spec.webhook.httpConfig.tlsConfig.minVersion
            string
            Minimum TLS version.
          • spec.webhook.httpConfig.tlsConfig.serverName
            string
            Server name for TLS certificate verification (SNI).
      • spec.webhook.url
        string

        Required value

        Webhook URL.

        The module sends notifications to this URL as JSON in the Alertmanager webhook format.

        Minimum length: 1

    • spec.zabbix
      object
      Parameters of the Zabbix channel.
      • spec.zabbix.api
        object

        Connection to the Zabbix API for automatic import of the template.

        If the section is set, the module imports the template and creates the cluster host and host group through the Zabbix JSON-RPC API. Otherwise, import the template from the ConfigMap manually.

        • spec.zabbix.api.tlsConfig
          object
          TLS settings of the HTTPS connection to the Zabbix API. Certificates and keys are given inline as PEM.
          • spec.zabbix.api.tlsConfig.ca
            string
            PEM-encoded CA certificate(s) used to verify the Zabbix API server certificate. If unset, the system trust store is used.
          • spec.zabbix.api.tlsConfig.caFile
            Deprecated
            string
            Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into ca instead.
          • spec.zabbix.api.tlsConfig.cert
            string
            PEM-encoded client certificate, for an API behind mutual TLS. Must be set together with key.
          • spec.zabbix.api.tlsConfig.certFile
            Deprecated
            string
            Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into cert instead.
          • spec.zabbix.api.tlsConfig.insecureSkipVerify
            boolean
            Disable verification of the Zabbix API server TLS certificate.
          • spec.zabbix.api.tlsConfig.key
            string
            PEM-encoded unencrypted private key for cert. Must be set together with cert.
          • spec.zabbix.api.tlsConfig.keyFile
            Deprecated
            string
            Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into key instead.
          • spec.zabbix.api.tlsConfig.serverName
            string
            Server name used for TLS certificate verification (SNI). Defaults to the host of url.
        • spec.zabbix.api.token
          string

          Required value

          Zabbix API token used to authenticate configuration.import calls.

          Minimum length: 1

        • spec.zabbix.api.url
          string

          Required value

          URL of the Zabbix JSON-RPC API, for example, https://zabbix.example.com/api_jsonrpc.php.

          If the scheme is omitted, https:// is used.

          Minimum length: 1

      • spec.zabbix.clusterName
        string

        Required value

        Name of the cluster in Zabbix.

        The name becomes the technical name of the Zabbix host that receives the data. It must be unique among all clusters and channels that send data to the same Zabbix instance.

        Minimum length: 1

      • spec.zabbix.disableTLS
        boolean

        Connect to the trapper port without TLS.

        If the parameter is disabled, the connection uses TLS with the settings from tls. If tls is omitted, the certificate is verified.

        Default: false

      • spec.zabbix.hostGroup
        string
        Zabbix host group that the cluster host is added to.

        Default: Deckhouse

      • spec.zabbix.port
        integer
        Trapper port of the Zabbix Server or Zabbix Proxy.

        Default: 10051

      • spec.zabbix.server
        string

        Required value

        Address of the Zabbix Server or Zabbix Proxy that receives trapper data.

        Minimum length: 1

      • spec.zabbix.tls
        object

        TLS settings of the connection to the Zabbix Server over the sender protocol.

        Certificates and keys are given inline as PEM. Zabbix certificate mode (TLSAccept=cert) is mutual TLS: cert and key are mandatory for it. Zabbix PSK mode is not supported.

        • spec.zabbix.tls.ca
          string
          PEM-encoded CA certificate(s) used to verify the Zabbix Server/Proxy certificate. If unset, the system trust store is used.
        • spec.zabbix.tls.caFile
          Deprecated
          string
          Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into ca instead.
        • spec.zabbix.tls.cert
          string
          PEM-encoded client certificate presented to the Zabbix Server/Proxy. Must be set together with key.
        • spec.zabbix.tls.certFile
          Deprecated
          string
          Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into cert instead.
        • spec.zabbix.tls.insecureSkipVerify
          boolean
          Disable verification of the Zabbix server TLS certificate entirely. Does not lift the client certificate requirement of Zabbix certificate mode.
        • spec.zabbix.tls.key
          string
          PEM-encoded unencrypted private key for cert. Must be set together with cert.
        • spec.zabbix.tls.keyFile
          Deprecated
          string
          Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into key instead.
        • spec.zabbix.tls.serverName
          string
          Server name used for TLS certificate verification (SNI). Defaults to spec.zabbix.server.
        • spec.zabbix.tls.skipServerNameVerification
          boolean

          Verify the server certificate chain against ca but do not match the certificate against serverName.

          Needed when the Zabbix certificate has no Subject Alternative Name (Zabbix validates only the Subject, so this is common), which is otherwise rejected. Requires ca: against the system trust store alone, a chain check without a name would accept any publicly trusted certificate. Ignored when insecureSkipVerify is true.

  • status
    object
    Status of the notification channel, including which policies are using it.
    • status.policies
      array of objects
      List of notification policies that are using this channel across all namespaces.
      • status.policies.name
        string
        Name of the notification policy.
      • status.policies.namespace
        string

        Namespace of the notification policy.

        Empty string for ClusterObservabilityPropagatedNotificationPolicy.

ClusterObservabilityPropagatedNotificationPolicy

Scope: Cluster
Version: v1alpha1

Rules for sending notifications about alerts that apply in every namespace.

Using the policy, the module selects alerts of each namespace by their labels and sends notifications about them to a ClusterObservabilityPropagatedNotificationChannel. In each namespace, the policy works the same way as an ObservabilityNotificationPolicy created in it.

  • spec
    object
    Policy parameters.
    • spec.alert
      object

      Required value

      Alerts that the policy applies to.
      • spec.alert.groupByLabels
        array of strings

        Labels by which alerts are grouped into one notification.

        Alerts with the same values of these labels are sent in one message. If the parameter is omitted, a separate notification is sent for each alert.

        • spec.alert.groupByLabels.Element of the array
          string
          Alert label name.
      • spec.alert.selector
        object

        Required value

        Selector of the alerts that the policy applies to.

        Set only one of the fields: matchLabels or matchExpressions.

        • spec.alert.selector.matchExpressions
          array of objects

          Set of expressions (conditions) for filtering alerts.

          The policy applies to an alert if the alert meets all of the conditions.

          • spec.alert.selector.matchExpressions.key
            string
            Alert label name.
          • spec.alert.selector.matchExpressions.operator
            string

            Comparison operator:

            • Equal: The label value equals value.
            • NotEqual: The label value differs from value.
            • Regex: The label value matches the value regular expression.
            • NotRegex: The label value does not match the value regular expression.

            Allowed values: Equal, NotEqual, Regex, NotRegex

          • spec.alert.selector.matchExpressions.value
            string
            Value or regular expression to compare the label value with.
        • spec.alert.selector.matchLabels
          object
          Labels and their values that an alert must have for the policy to apply to it.
    • spec.notification
      object

      Required value

      Notification parameters.
      • spec.notification.channel
        object

        Required value

        “Channel that receives the notifications.”
        • spec.notification.channel.kind
          string

          Required value

          Kind of the notification channel object.

          Default: ClusterObservabilityPropagatedNotificationChannel

          Allowed values: ClusterObservabilityPropagatedNotificationChannel

        • spec.notification.channel.name
          string

          Required value

          Name of the channel object.
        • spec.notification.channel.parameters
          object

          Required value

          Notification channel parameters.
      • spec.notification.repeatInterval
        string

        Required value

        Interval for resending the notification about an alert.

        Default: 1m

  • status
    object
    Notification policy state.
    • status.description
      string
      Reason why the module does not apply the policy.
    • status.state
      string

      Policy state:

      • Applicable: The module sends notifications according to the policy.
      • NotApplicable: The module does not apply the policy, for example, because the channel does not exist or parameters do not match the channel type. The reason is in the description field.

ObservabilityDashboard

Scope: Namespaced
Version: v1alpha1

Dashboard of a single namespace in the Deckhouse web UI. The dashboard is displayed in the project of the namespace where the object is created. To configure how the dashboard is displayed, use the following annotations:

  • metadata.deckhouse.io/category: Folder (category) of the dashboard in the interface. - metadata.deckhouse.io/title: Title of the dashboard in the interface. If the annotation is omitted, the module uses the title from the JSON manifest.
  • spec
    object
    • spec.definition
      string

      Required value

      Dashboard JSON manifest as a string compatible with the Grafana format.

ObservabilityDiagnosticReport

Short names: odr

Scope: Namespaced
Version: v1alpha1

Experimental. Diagnostic report of an alert of the namespace (ObservabilityAlert).

The controller creates the report when the alert fires: it finds the object of the alert, collects facts and evaluates the findings of the playbooks. A user with edit rights can create a report with only spec.alert.kind, spec.alert.name and spec.alert.namespace set to run the diagnostics manually, or request a new run of an existing report with the annotation observability.deckhouse.io/rerun.

Only the controller writes status.

  • spec
    object
    Defines the report. Immutable once the controller filled the alert snapshot.
    • spec.alert
      object

      Required value

      Reference to the alert object and the snapshot of the alert. A user creating a report sets only kind, name and namespace; the controller fills the snapshot.
      • spec.alert.alertName
        string
        Name of the alert.

        Maximum length: 253

      • spec.alert.annotations
        object
        Allowlisted annotations of the alert, each at most 1 KiB and at most 4 KiB together; the controller cuts a longer annotation and ends it with [truncated].
      • spec.alert.createdBy
        string
        Who created the rule of the alert.

        Maximum length: 253

      • spec.alert.fingerprint
        string
        Fingerprint of the alert.

        Maximum length: 64

      • spec.alert.kind
        string

        Required value

        Kind of the alert object.

        Allowed values: ObservabilityAlert

      • spec.alert.labels
        object
        Labels of the alert, at most 2 KiB together with their names.
      • spec.alert.name
        string

        Required value

        Name of the alert object.

        Length: 1..253

      • spec.alert.namespace
        string

        Required value

        Namespace of the alert object; for ObservabilityDiagnosticReport the namespace of the report.

        Maximum length: 63

      • spec.alert.rule
        object
        Rule context of the alert.
        • spec.alert.rule.description
          string
          Description template of the rule when it differs from the annotation.

          Maximum length: 2048

        • spec.alert.rule.expr
          string
          Expression of the rule.

          Maximum length: 2048

        • spec.alert.rule.for
          string
          for of the rule.

          Maximum length: 32

        • spec.alert.rule.keepFiringFor
          string
          keepFiringFor of the rule.

          Maximum length: 32

        • spec.alert.rule.summary
          string
          Summary template of the rule when it differs from the annotation.

          Maximum length: 1024

      • spec.alert.ruleGroupName
        string
        Rule group of the alert.

        Maximum length: 253

      • spec.alert.severity
        string
        Severity level of the alert.

        Maximum length: 16

      • spec.alert.startsAt
        string
        Time the alert started firing.
  • status
    object
    Written by the controller only.
    • status.alertState
      object
      State of the alert.
      • status.alertState.lostAt
        string
        When the report became Lost: once the alert object has stayed missing for the grace of the controller (3 minutes by default) without resolving.
      • status.alertState.resolvedAt
        string
        When the alert resolved.
      • status.alertState.state
        string
        State.

        Allowed values: Firing, Resolved, Lost

    • status.checks
      array of objects
      Results of the checks of the current run.
      • status.checks.coverage
        object
        How many objects of the population were checked in detail.
        • status.checks.coverage.checked
          integer

          Required value

          Checked in detail.

          Allowed values: 0 <= X

        • status.checks.coverage.total
          integer

          Required value

          Population.

          Allowed values: 0 <= X

      • status.checks.evidence
        object
        Allowlisted projection of the read data; the shape depends on the check type.
      • status.checks.evidenceKind
        string
        Name of the schema of evidence, one per shape (Events, PodLogs, Metrics, Deployment, WorkloadPods, Omitted when the budget of the run replaced the evidence). Absent on a row without evidence and in a report written before the field existed.

        Maximum length: 64

        Pattern: ^[A-Z][A-Za-z0-9]*$

      • status.checks.hidden
        boolean
        The row is not listed to a reader - a repeat with sameAs, or the snapshot of the rule baseline/alert-context.
      • status.checks.id
        string
        baseline/<check>, baseline/<kind>-<name>/<check> with several targets (baseline/<kind>-<namespace>-<name>/<check> when another target has its kind and name, with -<hash> after the name when two targets would share a prefix), baseline/<n>-<kind>-<name>/<check> of a linked object, or <playbook>/<check>.

        Maximum length: 320

      • status.checks.objectRef
        object
        Object the check read.
        • status.checks.objectRef.apiVersion
          string
          API version.

          Maximum length: 128

        • status.checks.objectRef.kind
          string

          Required value

          Kind.

          Maximum length: 63

        • status.checks.objectRef.name
          string

          Required value

          Name.

          Maximum length: 253

        • status.checks.objectRef.namespace
          string
          Namespace.

          Maximum length: 63

        • status.checks.objectRef.resourceVersion
          string
          Resource version at the time of reading.

          Maximum length: 64

        • status.checks.objectRef.uid
          string
          UID.

          Maximum length: 64

      • status.checks.observedAt
        string
        Time the fact was read.
      • status.checks.origin
        string
        Who planned the check.

        Allowed values: Baseline, Playbook

      • status.checks.outcome
        string
        Outcome. Anything other than Ok and ProblemFound means “unknown”, not “no problem”.

        Allowed values: Ok, ProblemFound, NotFound, InsufficientInput, Forbidden, Timeout, Error, Skipped

      • status.checks.reason
        string
        Stable CamelCase reason of the outcome.

        Maximum length: 63

      • status.checks.result
        string
        Verdict of the check: Problem - found a deviation, even in a partial read; Ok - a check with its own verdict found nothing; Partial - no deviation, but only part was read or the evidence was cut; NoData - nothing to judge (outcome InsufficientInput, NotFound, Skipped, or no metric sample in the window); Failed - no answer (outcome Error, Forbidden, Timeout); Context - data without a verdict. Absent in older reports.

        Allowed values: Problem, Ok, Partial, NoData, Failed, Context

      • status.checks.sameAs
        string
        ID of an earlier check of the run with the same canonical key; such a check stores no evidence.

        Maximum length: 320

      • status.checks.summary
        string
        One line for a human, in English.

        Maximum length: 300

      • status.checks.targetRef
        object
        Target of the run (an entry of status.run.targets) or linked object (an entry of status.run.linked) the row belongs to. Absent on the rows of the run as a whole (baseline/alert-context, baseline/linked-alerts) and in a report written before the field existed. objectRef stays the object the check read.
        • status.checks.targetRef.apiVersion
          string
          API version.

          Maximum length: 128

        • status.checks.targetRef.kind
          string

          Required value

          Kind.

          Maximum length: 63

        • status.checks.targetRef.name
          string

          Required value

          Name.

          Maximum length: 253

        • status.checks.targetRef.namespace
          string
          Namespace.

          Maximum length: 63

        • status.checks.targetRef.resourceVersion
          string
          Resource version at the time of reading.

          Maximum length: 64

        • status.checks.targetRef.uid
          string
          UID.

          Maximum length: 64

      • status.checks.title
        string
        Name of the check for a reader, in English, 1 to 4 words: of the table of the controller for a row of the baseline, spec.checks[].title of the playbook for a check of a playbook, else built from its type and subject. The controller writes it on every row; absent in a report written before the field existed.

        Maximum length: 64

      • status.checks.truncated
        boolean
        Evidence was cut by the limits; the summary names what was removed.
      • status.checks.type
        string
        Check type; NodeAgent is a call of the node agent of the node of a Node target, ScrapeTargets reads the Prometheus scrape targets through label-enforcer, Baseline is a check of a recipe of a playbook that did not run (its condition is false, a parameter failed or the object is linked), whose recipe wrote no row.

        Allowed values: Object, Related, List, Events, PodLogs, PodExec, Metrics, RelatedAlerts, WorkloadContext, ServiceBackends, AlertContext, NodeAgent, ScrapeTargets, Baseline

    • status.collection
      object
      Fact collection of the current run.
      • status.collection.checksDone
        integer
        Finished checks.

        Allowed values: 0 <= X

      • status.collection.checksTotal
        integer
        Planned checks.

        Allowed values: 0 <= X

      • status.collection.completedAt
        string
        End.
      • status.collection.message
        string
        Message for a human.

        Maximum length: 1024

      • status.collection.phase
        string
        Phase.

        Allowed values: Pending, Running, Completed, Partial, Failed

      • status.collection.reason
        string
        Reason of a failed collection.

        Maximum length: 63

      • status.collection.startedAt
        string
        Start.
    • status.conditions
      array of objects
      Conditions.
      • status.conditions.lastTransitionTime
        string
        Time of the last change of the status.
      • status.conditions.message
        string
        Message for a human.

        Maximum length: 32768

      • status.conditions.observedGeneration
        integer
        Generation the condition describes.

        Allowed values: 0 <= X

      • status.conditions.reason
        string
        Machine-readable reason.

        Maximum length: 1024

      • status.conditions.status
        string
        Status.

        Allowed values: True, False, Unknown

      • status.conditions.type
        string
        Type.

        Maximum length: 316

    • status.environment
      object
      Snapshot of versions.
      • status.environment.dkpVersion
        string
        Deckhouse version.

        Maximum length: 64

      • status.environment.kubernetesVersion
        string
        Kubernetes version.

        Maximum length: 64

      • status.environment.modules
        array of objects
        Modules the playbooks name.
        • status.environment.modules.enabled
          boolean
          Whether the module is enabled.
        • status.environment.modules.name
          string
          Name.

          Maximum length: 63

      • status.environment.observabilityVersion
        string
        Version of the module.

        Maximum length: 64

    • status.findings
      array of objects
      Findings of the current run.
      • status.findings.actions
        array of strings
        What to do, in Markdown.
        • status.findings.actions.Element of the array
          string

          Maximum length: 1024

      • status.findings.causedBy
        array of strings
        IDs of the findings of this run the finding follows from. Empty for a finding that stands on its own; such findings come first.
        • status.findings.causedBy.Element of the array
          string

          Maximum length: 128

      • status.findings.demotion
        string
        Why a behaviour cause was taken back: the sentence on the basic component that was not ready in the window of the alert. status.summary.causeGap repeats it when the finding gives the headline.

        Maximum length: 512

      • status.findings.evidenceRefs
        array of strings
        IDs of the checks the finding is based on.
        • status.findings.evidenceRefs.Element of the array
          string

          Maximum length: 320

      • status.findings.explanation
        string
        Why, in Markdown.

        Maximum length: 4096

      • status.findings.facts
        array of objects
        The values that prove a finding, in the order of its playbook, or of the file of the texts of the controller for a built-in finding: a fact whose value rendered empty, or whose expression failed, is not here. A built-in finding may carry facts too.
        • status.findings.facts.checkId
          string
          ID of the first check the expressions of the value read that the run holds, in the form of evidenceRefs: baseline/<check> with the prefix of the target or of the linked object, or <playbook>/<check>. Empty for a value that reads no check or none the run holds. For a fact of a built-in finding, the check its value was read from, else the first evidenceRefs of the finding.

          Maximum length: 320

        • status.findings.facts.highlight
          boolean
          The fact carries the problem itself, such as the text of an error, the reason of a refusal or the message of a component: its playbook, or the file of the texts of the controller for a built-in finding, marks it. A reader shows it first. At most 2 facts of a finding are marked; a fact whose value rendered empty is not here, with its mark.
        • status.findings.facts.inPrimaryEvidence
          boolean
          The fact only repeats what primaryEvidence of the finding shows, such as the largest directory of a tree: a reader that draws that evidence with the finding may leave the fact out. Set only when the finding has primaryEvidence.
        • status.findings.facts.label
          string
          What the value is, as the playbook names it.

          Maximum length: 64

        • status.findings.facts.value
          string
          The value in one line.

          Maximum length: 256

      • status.findings.id
        string
        <playbook>/<finding>, baseline/<finding> for a built-in finding of the controller, or baseline/<n>-<kind>-<name>/<finding> for a built-in finding of a linked object.

        Maximum length: 128

      • status.findings.objectRef
        object
        Object the finding is about: the target of its playbook run or of the prefix of its id (an entry of status.run.targets), or the linked object of its playbook run or of its prefix (an entry of status.run.linked). Absent on a built-in finding about no single object and in a report written before the field existed.
        • status.findings.objectRef.apiVersion
          string
          API version.

          Maximum length: 128

        • status.findings.objectRef.kind
          string

          Required value

          Kind.

          Maximum length: 63

        • status.findings.objectRef.name
          string

          Required value

          Name.

          Maximum length: 253

        • status.findings.objectRef.namespace
          string
          Namespace.

          Maximum length: 63

        • status.findings.objectRef.resourceVersion
          string
          Resource version at the time of reading.

          Maximum length: 64

        • status.findings.objectRef.uid
          string
          UID.

          Maximum length: 64

      • status.findings.primaryEvidence
        string
        ID of the check whose evidence proves the finding, in the form of evidenceRefs: the primaryEvidence of its playbook, set only when the run holds that check. The console draws that evidence with the finding. Absent for a built-in finding and when the playbook names none.

        Maximum length: 320

      • status.findings.references
        array of objects
        Up to 3 links of the finding to a Deckhouse module, in the order of its playbook. The console builds the URL: ModuleSetting - the module settings and the parameter in its documentation; ModuleResource - the resource in the module documentation; ModuleDocs - a documentation page. Never a URL or a setting value; a console that does not know a kind skips it. Absent in older reports.
        • status.findings.references.anchor
          string
          ModuleDocs only: the anchor on the page, as in the published documentation.

          Maximum length: 128

          Pattern: ^[\p{L}\p{N}_][\p{L}\p{N}_.-]*$

        • status.findings.references.kind
          string
          What the reference points to.

          Allowed values: ModuleSetting, ModuleResource, ModuleDocs

        • status.findings.references.module
          string
          The module by its name in Module and ModuleConfig; global for the global ModuleConfig.

          Maximum length: 64

          Pattern: ^[a-z0-9]([-a-z0-9]*[a-z0-9])?$

        • status.findings.references.page
          string
          ModuleDocs only: the page of the documentation of the module.

          Allowed values: Overview, Configuration, Resources, Usage, FAQ, Examples

        • status.findings.references.resourceKind
          string
          ModuleResource only: the kind of a resource of the module.

          Maximum length: 63

          Pattern: ^[A-Z][A-Za-z0-9]*$

        • status.findings.references.resourceVersion
          string
          ModuleResource only: the version of the API of the resource, such as v1; absent when it is not known.

          Maximum length: 32

          Pattern: ^v\d+((alpha|beta)\d+)?$

        • status.findings.references.setting
          string
          ModuleSetting only: the dot-separated path of the key in the settings of the module, without array indexes. Never a value.

          Maximum length: 256

          Pattern: ^[A-Za-z_][A-Za-z0-9_-]*(\.[A-Za-z_][A-Za-z0-9_-]*)*$

      • status.findings.role
        string
        What the finding says about the cause of the alert: Cause, ConfigurationCause or Symptom as set in the playbook (a withdrawn behaviour cause is Symptom); Fact for a built-in fact; ObjectProblem for a would-be cause found by a fallback playbook: a problem of the object from the alert labels, not the cause of the alert. symptom is true exactly for Symptom.

        Allowed values: Cause, ConfigurationCause, Symptom, Fact, ObjectProblem

      • status.findings.severity
        string
        Severity.

        Allowed values: Info, Warning, Critical

      • status.findings.symptom
        boolean
        The finding only restates what the alert or the object shows and does not establish the root cause by itself. The marker does not change causedBy.
      • status.findings.title
        string
        Title, one line.

        Maximum length: 200

    • status.history
      array of objects
      Up to five past runs.
      • status.history.completedAt
        string
        End.
      • status.history.findings
        array of objects
        Findings.
        • status.history.findings.evidenceRefs
          array of strings
          IDs of the checks.
          • status.history.findings.evidenceRefs.Element of the array
            string

            Maximum length: 320

        • status.history.findings.id
          string
          Finding ID.

          Maximum length: 128

        • status.history.findings.severity
          string
          Severity.

          Allowed values: Info, Warning, Critical

        • status.history.findings.title
          string
          Title.

          Maximum length: 200

      • status.history.headline
        string
        Headline.

        Maximum length: 300

      • status.history.id
        string
        Run ID.

        Maximum length: 64

      • status.history.startedAt
        string
        Start.
      • status.history.trigger
        string
        Trigger.

        Allowed values: Alert, Manual

    • status.platformNotes
      array of objects
      One-line notes under the findings. For a target of class Platform, first the findings of a later hypothesis step that were not taken as the cause, one per finding. Then the basic components on the path of a target that were not ready during the alert but did not withdraw a cause, because their failure began after the alert or at an unknown time; one per component and target.
      • status.platformNotes.checkId
        string
        ID of the platform-state check the note was read from; for a note of a hypothesis, the first evidence of the finding.

        Maximum length: 320

      • status.platformNotes.component
        string
        Name of the component; in a project report only the class of the failure. A note of a hypothesis names the module of the target, else the component of the playbook of the finding, else platform.

        Maximum length: 63

      • status.platformNotes.message
        string
        One sentence: what was not ready, the source it was read from, and that its failure started after the alert or that its start is not known. A project report names only the class of the failure. A note of a hypothesis gives the title of the finding and the title of the cause that comes first in the order of the hypotheses.

        Maximum length: 512

      • status.platformNotes.report
        string
        Name of the cluster report of the firing alert of the component; never in a project report.

        Maximum length: 253

      • status.platformNotes.since
        string
        The earliest moment the source proves the component not ready; absent when none is proven.
    • status.rejectedRerunNonce
      string
      Last rejected value of the annotation observability.deckhouse.io/rerun.

      Maximum length: 256

    • status.run
      object
      Current run.
      • status.run.attempt
        integer

        Required value

        Execution of the run; a new leader repeats an unfinished run with the next attempt, at most 3.

        Allowed values: 0 <= X

      • status.run.id
        string

        Required value

        Run ID.

        Maximum length: 64

      • status.run.linked
        array of objects
        Objects the run found from its targets, in the order they were found. The ids of the checks and built-in findings of such an object start with its checkPrefix; their outcomes do not change the phase of the run.
        • status.run.linked.apiVersion
          string
          API version.

          Maximum length: 128

        • status.run.linked.checkPrefix
          string
          Prefix of the ids of its checks and built-in findings: baseline/<n>-<kind>-<name>/, n is the order of the object, the name is cut to 40 characters.

          Maximum length: 320

        • status.run.linked.class
          string
          Platform for an object of Deckhouse: the label heritage: deckhouse on the object or on its namespace, a namespace d8-* or kube-system, a Node, or a cluster resource of an API group of Deckhouse; User for any other object. Written for a linked object that was read. Absent in a report written before the field existed.

          Allowed values: Platform, User

        • status.run.linked.kind
          string
          Kind.

          Maximum length: 63

        • status.run.linked.module
          string
          Module of Deckhouse of an object of class Platform, the label module of the object. Absent for class User and when neither names a module.

          Maximum length: 64

        • status.run.linked.moduleVersion
          string
          properties.version of the Module named by module. Absent when the Module was not found or could not be read, when the scope of the report does not read the Module (a project report whose class Platform rests on the labels of the object alone), and when the value is longer than 64 characters.

          Maximum length: 64

        • status.run.linked.name
          string
          Name.

          Maximum length: 253

        • status.run.linked.namespace
          string
          Namespace.

          Maximum length: 63

        • status.run.linked.of
          string
          checkPrefix of the object it was found from: baseline/ or baseline/<kind>-<name>/ of a target (baseline/<kind>-<namespace>-<name>/ when another target has its kind and name, with -<hash> at the end when two targets would share a prefix), or of another linked object.

          Maximum length: 320

        • status.run.linked.parentRef
          object
          Object it was found from, the target (an entry of status.run.targets) or the linked object (an entry of status.run.linked) whose prefix is of. Absent in a report written before the field existed.
          • status.run.linked.parentRef.apiVersion
            string
            API version.

            Maximum length: 128

          • status.run.linked.parentRef.kind
            string

            Required value

            Kind.

            Maximum length: 63

          • status.run.linked.parentRef.name
            string

            Required value

            Name.

            Maximum length: 253

          • status.run.linked.parentRef.namespace
            string
            Namespace.

            Maximum length: 63

          • status.run.linked.parentRef.resourceVersion
            string
            Resource version at the time of reading.

            Maximum length: 64

          • status.run.linked.parentRef.uid
            string
            UID.

            Maximum length: 64

        • status.run.linked.relation
          string
          Backend: the workload or the pod behind the Service an Ingress target routes to. Dependency: the object behind an address a log line connects to.

          Allowed values: Backend, Dependency

        • status.run.linked.stateCheckId
          string
          ID in status.checks of the row that holds the state of the object: <checkPrefix>target, written with the rows of the linked phase. The row is there also when the object was not found (its outcome tells). Absent when the run wrote no such row (the time or the checks of the run ran out before the object was read) and in a report written before the field existed.

          Maximum length: 320

        • status.run.linked.stoppedBy
          string
          Why the branch does not go on from the object; empty when it goes on.

          Allowed values: NoAddress, External, OutOfScope, Loop, MaxDepth, MaxObjects, TimeSlice, NotFound, MaxChecks, NotRead

        • status.run.linked.via
          object
          What the object was found through.
          • status.run.linked.via.check
            string

            Required value

            ID of the check of the object before it that named it.

            Maximum length: 320

          • status.run.linked.via.declaredIn
            string
            Where the configuration names the address: <container>: env <NAME> or <container>: args; the value is never stored.

            Maximum length: 384

          • status.run.linked.via.host
            string
            Host of the address of a log line, for a Dependency.

            Maximum length: 253

          • status.run.linked.via.port
            string
            Port of the address of a log line or of the configuration, for a Dependency.

            Maximum length: 5

          • status.run.linked.via.service
            string
            Service whose pods the object runs.

            Maximum length: 63

          • status.run.linked.via.source
            string
            Webhook for a Backend found through the configuration of a webhook; Config for a Dependency found by the configuration of a container (the literal values of its env, its command and args); omitted for one found by a log line.

            Allowed values: Config, Webhook

      • status.run.owner
        string
        Controller pod that executes the run.

        Maximum length: 253

      • status.run.playbooks
        array of objects
        Playbooks that took part in the run.
        • status.run.playbooks.fallback
          boolean
          true for a built-in playbook that ran on the object the labels of the alert name because no playbook names the alert; its findings describe that object, and a finding that would name a cause has the role ObjectProblem. Absent for a playbook of the alert and in a report written before the field existed.
        • status.run.playbooks.generation
          integer
          Generation.

          Allowed values: 0 <= X

        • status.run.playbooks.name
          string
          Name.

          Maximum length: 253

        • status.run.playbooks.source
          string
          Where the playbook comes from - Builtin, the image of the controller with generation 0, or Cluster, a ClusterObservabilityDiagnosticPlaybook.

          Allowed values: Builtin, Cluster

      • status.run.rerunNonce
        string
        Accepted value of the annotation observability.deckhouse.io/rerun.

        Maximum length: 256

      • status.run.startedAt
        string

        Required value

        Start of the run.
      • status.run.targetSource
        string
        Where the targets come from - Playbook, the target a playbook of the alert addresses, or AlertLabels, the object the labels of an alert that no playbook matches name, also when a playbook runs on it as a fallback. Empty for a run without a target.

        Allowed values: Playbook, AlertLabels

      • status.run.targets
        array of objects
        Diagnosed objects.
        • status.run.targets.apiVersion
          string
          API version.

          Maximum length: 128

        • status.run.targets.class
          string
          Platform for a Deckhouse object (label heritage: deckhouse on the object or its namespace, a d8-* or kube-system namespace, a Node, a cluster resource of a Deckhouse API group); User otherwise. When the object cannot be read, its namespace decides. Absent in older reports.

          Allowed values: Platform, User

        • status.run.targets.kind
          string
          Kind.

          Maximum length: 63

        • status.run.targets.module
          string
          Module of Deckhouse of an object of class Platform, the module of the component of its playbook, else the label module of the object, sanitized as the evidence is. Absent for class User, when neither names a module, and when the value is longer than 64 characters.

          Maximum length: 64

        • status.run.targets.moduleVersion
          string
          properties.version of the Module named by module. Absent when the Module was not found or could not be read, when the scope of the report does not read the Module (a project report whose class Platform rests on the labels of the object alone), and when the value is longer than 64 characters.

          Maximum length: 64

        • status.run.targets.name
          string
          Name.

          Maximum length: 253

        • status.run.targets.namespace
          string
          Namespace.

          Maximum length: 63

        • status.run.targets.resourceVersion
          string
          Resource version at the time of reading.

          Maximum length: 64

        • status.run.targets.stateCheckId
          string
          ID in status.checks of the row with the state of the object: baseline/target, or baseline/<kind>-<name>/target with several targets; present also when the object was not found (see its outcome). Absent until the row is written, when every playbook sets baseline: None, and in older reports.

          Maximum length: 320

        • status.run.targets.uid
          string
          UID.

          Maximum length: 64

      • status.run.trigger
        string

        Required value

        Alert for an automatic run, Manual for a report created by a user or a rerun.

        Allowed values: Alert, Manual

    • status.summary
      object
      Block for the UI.
      • status.summary.busy
        boolean
        A run is in progress.
      • status.summary.causeEstablished
        boolean
        true only when the headline is a finding that names a cause: not a symptom, not a fact and not a problem of the object (ObjectProblem), and Warning or Critical or of the role ConfigurationCause.
      • status.summary.causeGap
        string
        With causeEstablished false, one sentence in English on what the report could not read or follow.

        Maximum length: 300

      • status.summary.currentActivity
        string
        What the controller does or did, in English.

        Maximum length: 200

      • status.summary.findingsCount
        integer
        Number of findings.

        Allowed values: 0 <= X

      • status.summary.headline
        string
        Title of the main finding: the most severe one without causedBy that names a cause, else the most severe one without causedBy, else the most severe one; never a built-in fact. Starts with Cause not established: when it names no cause, e.g. Cause not established: no playbook matches the alert; with a fallback playbook, No playbook matches the alert: and the title of an ObjectProblem. Empty while the report is collecting.

        Maximum length: 300

      • status.summary.headlineFindingId
        string
        ID of the finding the headline comes from. Empty when no finding gives the headline.

        Maximum length: 320

      • status.summary.headlineSource
        string
        Finding when a finding gives the headline; empty otherwise.

        Allowed values: Finding

      • status.summary.progress
        object
        Progress of the collection.
        • status.summary.progress.checksDone
          integer
          Finished checks.

          Allowed values: 0 <= X

        • status.summary.progress.checksTotal
          integer
          Planned checks.

          Allowed values: 0 <= X

      • status.summary.severity
        string
        Highest severity of the findings.

        Allowed values: Info, Warning, Critical

      • status.summary.state
        string
        Collecting while the run goes, Ready when it is final, Failed when facts could not be collected. Derived from collection.phase.

        Allowed values: Collecting, Ready, Failed

      • status.summary.updatedAt
        string
        Time of the last update.

ObservabilityLogsRulesGroup

Short names: olrg

Scope: Namespaced
Version: v1alpha1

Group of log-based recording rules of the namespace scope.

Rules are evaluated by the logs-ruler service, which makes LogQL queries to the log storage backend. The resulting series are exposed as regular Prometheus metrics.

To set up an alert based on the created recording rules, use the ObservabilityMetricsRulesGroup resource.

LogQL expressions are evaluated for logs of a namespace where the resource is created.

  • spec
    object
    Parameters of a group of the log-based recording rules.
    • spec.interval
      string

      Rule evaluation interval.

      For example, 30s or 1m.

      If not specified, the default value is used.

      Default: 1m

      Pattern: ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$

    • spec.rules
      array of objects

      Required value

      List of log-based recording rules.
      • spec.rules.expr
        string

        LogQL expression to evaluate the recording rule.

        Do not use the namespace label in the expression. The module adds it automatically based on the resource namespace.

        Minimum length: 1

      • spec.rules.labels
        object

        Additional labels attached to the recorded metric series.

        The following labels are reserved and can’t be set by the user:

        • namespace
        • scope_type
        • created_by
        • rule_group_name
      • spec.rules.record
        string

        Name of the recording rule.

        Must comply with the Prometheus metric naming rules and be unique within the rule group.

        Minimum length: 1

        Pattern: ^[a-zA-Z_:][a-zA-Z0-9_:]*$

ObservabilityMetricsRulesGroup

Short names: omrg

Scope: Namespaced
Version: v1alpha1

Group of alerting or recording rules for a single namespace.
  • spec
    object
    Rule group parameters.
    • spec.interval
      string

      Group rule evaluation interval.

      If not specified, the default interval is used.

      Default: 1m

      Pattern: ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$

    • spec.rules
      array of objects

      Required value

      List of alerting and recording rules of the group.
      • spec.rules.alert
        string

        Name of the alerting rule.

        Only one of the fields can be specified at the same time: alert or record.

        Pattern: ^[a-zA-Z-:][a-zA-Z0-9-:]*$

      • spec.rules.annotations
        object

        Annotations to attach to the alert.

        Applies only to alerting rules.

      • spec.rules.expr
        integer or string
        PromQL expression to evaluate the rule.
      • spec.rules.for
        string

        Delay before the alert is triggered.

        Defines how long the condition must continuously be true before the alert transitions to the firing state.

        Applies only to alerting rules.

        Pattern: ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$

      • spec.rules.keep_firing_for
        string

        Delay before the alert is resolved.

        Defines how long the alert remains in the firing state after the condition becomes false.

        Applies only to alerting rules.

        Pattern: ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$

      • spec.rules.labels
        object
        Additional labels to attach to the alert or recording rule.
      • spec.rules.record
        string

        Name of the recording rule.

        Only one of the fields can be specified at the same time: alert or record.

        Pattern: ^[a-zA-Z_:][a-zA-Z0-9_:]*$

  • status
    object
    Rule group state.
    • status.conditions
      array of objects
      The latest available observed state of the rules group.
      • status.conditions.lastTransitionTime
        string
        Last time the condition transitioned from one status to another.
      • status.conditions.message
        string
        Human-readable message indicating details about last transition.
      • status.conditions.reason
        string

        Reason for the condition’s last transition.

        Specified in one word in CamelCase format.

      • status.conditions.status
        string

        Current status of the condition:

        • True: Condition is satisfied.
        • False: Condition is not satisfied.
        • Unknown: Condition could not be determined.

        Allowed values: True, False, Unknown

      • status.conditions.type
        string

        Type of the rules group condition:

        • Created: Resource is created.
        • Failed: Syncing failed.
        • Ready: Rules group is applied and active.

        Allowed values: Created, Failed, Ready

    • status.lastSyncTime
      string

      Last time the rules group was successfully synced, which changed its PrometheusRule or its status.

      Not refreshed when a sync changes nothing.

    • status.prometheusRuleName
      string
      Name of the PrometheusRule resource created for this rules group.

ObservabilityNotificationChannel

Scope: Namespaced
Version: v1alpha1

Notification channel for the alerts of one namespace.

The module sends a notification when an alert fires and when it is resolved. Notification rules are set by an ObservabilityNotificationPolicy object in the same namespace.

  • spec
    object

    Channel parameters.

    Set type and the section with the parameters of this type, for example, type: Email and email.

    • spec.email
      object
      Parameters of the Email channel.
      • spec.email.from
        string

        Required value

        Sender email address.
      • spec.email.smtp
        object

        Required value

        Parameters of the SMTP server that sends the emails.
        • spec.email.smtp.address
          string

          Required value

          Address of the SMTP server in the host:port format, for example, smtp.example.com:587.
        • spec.email.smtp.auth
          object
          Credentials for authentication on the SMTP server.
          • spec.email.smtp.auth.password
            string

            Required value

            Password.
          • spec.email.smtp.auth.username
            string

            Required value

            Username.
        • spec.email.smtp.requireTLS
          boolean
          Whether TLS verification is required.
      • spec.email.template
        string
        Email template.
    • spec.expressMessenger
      object
      Parameters of the ExpressMessenger channel.
      • spec.expressMessenger.apiUrl
        string

        Required value

        URL of the eXpress CTS server, for example, https://cts.company.ru.
      • spec.expressMessenger.botId
        string

        Required value

        ID (UUID) of the eXpress bot that sends the notifications.

        Pattern: ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$

      • spec.expressMessenger.secretKey
        string

        Required value

        Secret key of the bot for authentication (JWT v2).
      • spec.expressMessenger.template
        string
        Message template.
    • spec.slack
      object
      Parameters of the Slack channel.
      • spec.slack.apiUrl
        string

        Required value

        Slack API URL.
      • spec.slack.color
        string
        Color of the message attachment.
      • spec.slack.fallback
        string
        Fallback text of the message.
      • string
        Footer of the message.
      • spec.slack.iconEmoji
        string
        Emoji for the message.
      • spec.slack.iconUrl
        string
        URL of the icon image in the message.
      • spec.slack.pretext
        string
        Text above the main block of the message.
      • spec.slack.template
        string
        Message template.
      • spec.slack.text
        string
        Main text of the message.
      • spec.slack.title
        string
        Title of the message.
      • string
        URL that the title links to.
      • spec.slack.username
        string
        Name of the message sender.
    • spec.telegram
      object
      Parameters of the Telegram channel.
      • spec.telegram.apiToken
        string

        Required value

        API token of the Telegram bot that sends the notifications.
      • spec.telegram.apiUrl
        string
        Telegram Bot API URL.

        Default: https://api.telegram.org

      • spec.telegram.httpConfig
        object

        HTTP client parameters for requests to the Telegram API.

        Use proxyUrl or proxyFromEnvironment if the Telegram API is reachable only through a proxy.

        Set only one of the basicAuth, authorization and oauth2 parameters.

        • spec.telegram.httpConfig.authorization
          object
          Value of the Authorization header.
          • spec.telegram.httpConfig.authorization.credentials
            string
            Credentials inserted into the Authorization header.
          • spec.telegram.httpConfig.authorization.credentialsFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use credentials.

          • spec.telegram.httpConfig.authorization.type
            string

            Authentication scheme, for example, Bearer.

            If the parameter is omitted, Bearer is used.

        • spec.telegram.httpConfig.basicAuth
          object
          HTTP Basic authentication.
          • spec.telegram.httpConfig.basicAuth.password
            string
            Password.
          • spec.telegram.httpConfig.basicAuth.passwordFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use password.

          • spec.telegram.httpConfig.basicAuth.username
            string
            Username.
        • spec.telegram.httpConfig.enableHttp2
          boolean
          Enable HTTP/2 support.
        • spec.telegram.httpConfig.followRedirects
          boolean

          Deprecated and ignored for security reasons.

          The module does not follow redirects because a redirect makes Alertmanager send a request to an address chosen by the notification recipient.

        • spec.telegram.httpConfig.httpHeaders
          object

          Custom HTTP headers of requests to the Telegram API.

          The object key is the header name, the value is an object with a list of values and/or references to secrets and files.

          • spec.telegram.httpConfig.httpHeaders.<KEY_NAME>
            object
            <KEY_NAME> — item (key) name.
            • spec.telegram.httpConfig.httpHeaders.<KEY_NAME>.files
              array of strings

              Deprecated and ignored for security reasons.

              File contents are not inserted into request headers. Use values or secrets.

            • spec.telegram.httpConfig.httpHeaders.<KEY_NAME>.secrets
              array of strings
              Header values read from secrets.
            • spec.telegram.httpConfig.httpHeaders.<KEY_NAME>.values
              array of strings
              Header values.
        • spec.telegram.httpConfig.noProxy
          string
          Comma-separated list of hosts for which the proxy is not used.
        • spec.telegram.httpConfig.oauth2
          object
          OAuth 2.0 authentication.
          • spec.telegram.httpConfig.oauth2.clientId
            string

            Required value

            OAuth 2.0 client ID.
          • spec.telegram.httpConfig.oauth2.clientSecret
            string
            OAuth 2.0 client secret.
          • spec.telegram.httpConfig.oauth2.clientSecretFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use clientSecret.

          • spec.telegram.httpConfig.oauth2.endpointParams
            object
            Additional parameters passed in the token request.
          • spec.telegram.httpConfig.oauth2.noProxy
            string
            Comma-separated list of hosts for which the proxy is not used.
          • spec.telegram.httpConfig.oauth2.proxyFromEnvironment
            boolean
            Use the proxy from the HTTP_PROXY, HTTPS_PROXY and NO_PROXY environment variables.
          • spec.telegram.httpConfig.oauth2.proxyUrl
            string
            URL of the HTTP proxy used for the token request.
          • spec.telegram.httpConfig.oauth2.scopes
            array of strings
            Scopes requested for the token.
            • spec.telegram.httpConfig.oauth2.scopes.Element of the array
              string
              Scope name.
          • spec.telegram.httpConfig.oauth2.tlsConfig
            object
            TLS settings for the token request.
            • spec.telegram.httpConfig.oauth2.tlsConfig.caFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.telegram.httpConfig.oauth2.tlsConfig.certFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.telegram.httpConfig.oauth2.tlsConfig.insecureSkipVerify
              boolean
              Disable verification of the server TLS certificate.
            • spec.telegram.httpConfig.oauth2.tlsConfig.keyFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.telegram.httpConfig.oauth2.tlsConfig.maxVersion
              string
              Maximum TLS version.
            • spec.telegram.httpConfig.oauth2.tlsConfig.minVersion
              string
              Minimum TLS version.
            • spec.telegram.httpConfig.oauth2.tlsConfig.serverName
              string
              Server name for TLS certificate verification (SNI).
          • spec.telegram.httpConfig.oauth2.tokenUrl
            string

            Required value

            URL of the token endpoint.
        • spec.telegram.httpConfig.proxyConnectHeader
          object

          Headers of CONNECT requests to the proxy.

          Header names are mapped to lists of values.

        • spec.telegram.httpConfig.proxyFromEnvironment
          boolean
          Use the proxy from the HTTP_PROXY, HTTPS_PROXY and NO_PROXY environment variables.
        • spec.telegram.httpConfig.proxyUrl
          string
          URL of the HTTP proxy for requests to the Telegram API.
        • spec.telegram.httpConfig.tlsConfig
          object
          TLS settings for the HTTP client of the Telegram receiver.
          • spec.telegram.httpConfig.tlsConfig.caFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.telegram.httpConfig.tlsConfig.certFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.telegram.httpConfig.tlsConfig.insecureSkipVerify
            boolean
            Disable verification of the server TLS certificate.
          • spec.telegram.httpConfig.tlsConfig.keyFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.telegram.httpConfig.tlsConfig.maxVersion
            string
            Maximum TLS version.
          • spec.telegram.httpConfig.tlsConfig.minVersion
            string
            Minimum TLS version.
          • spec.telegram.httpConfig.tlsConfig.serverName
            string
            Server name for TLS certificate verification (SNI).
      • spec.telegram.template
        string
        Message template.
    • spec.type
      string

      Required value

      Notification channel type.

      Allowed values: Email, Slack, Telegram, Webhook, ExpressMessenger, Zabbix

    • spec.webhook
      object
      Parameters of the Webhook channel.
      • spec.webhook.httpConfig
        object

        HTTP client parameters for requests to the webhook URL.

        Set only one of the basicAuth, authorization and oauth2 parameters.

        • spec.webhook.httpConfig.authorization
          object
          Value of the Authorization header.
          • spec.webhook.httpConfig.authorization.credentials
            string
            Credentials inserted into the Authorization header.
          • spec.webhook.httpConfig.authorization.credentialsFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use credentials.

          • spec.webhook.httpConfig.authorization.type
            string

            Authentication scheme, for example, Bearer.

            If the parameter is omitted, Bearer is used.

        • spec.webhook.httpConfig.basicAuth
          object
          HTTP Basic authentication.
          • spec.webhook.httpConfig.basicAuth.password
            string
            Password.
          • spec.webhook.httpConfig.basicAuth.passwordFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use password.

          • spec.webhook.httpConfig.basicAuth.username
            string
            Username.
        • spec.webhook.httpConfig.enableHttp2
          boolean
          Enable HTTP/2 support for webhook requests.
        • spec.webhook.httpConfig.followRedirects
          boolean

          Deprecated and ignored for security reasons.

          The module does not follow HTTP redirects.

        • spec.webhook.httpConfig.httpHeaders
          object

          Custom HTTP headers of requests to the webhook URL.

          The object key is the header name, the value is an object with a list of values and/or references to secrets and files.

          • spec.webhook.httpConfig.httpHeaders.<KEY_NAME>
            object
            <KEY_NAME> — item (key) name.
            • spec.webhook.httpConfig.httpHeaders.<KEY_NAME>.files
              array of strings

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container. Use values or secrets.

            • spec.webhook.httpConfig.httpHeaders.<KEY_NAME>.secrets
              array of strings
              Header values read from secrets.
            • spec.webhook.httpConfig.httpHeaders.<KEY_NAME>.values
              array of strings
              Header values.
        • spec.webhook.httpConfig.noProxy
          string
          Comma-separated list of hosts that are accessed without the proxy.
        • spec.webhook.httpConfig.oauth2
          object
          OAuth 2.0 authentication.
          • spec.webhook.httpConfig.oauth2.clientId
            string

            Required value

            OAuth 2.0 client ID.
          • spec.webhook.httpConfig.oauth2.clientSecret
            string
            OAuth 2.0 client secret.
          • spec.webhook.httpConfig.oauth2.clientSecretFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container. Use clientSecret.

          • spec.webhook.httpConfig.oauth2.endpointParams
            object
            Additional parameters passed in the token request.
          • spec.webhook.httpConfig.oauth2.noProxy
            string
            Comma-separated list of hosts for which the proxy is not used.
          • spec.webhook.httpConfig.oauth2.proxyFromEnvironment
            boolean
            Use the proxy from the HTTP_PROXY, HTTPS_PROXY and NO_PROXY environment variables.
          • spec.webhook.httpConfig.oauth2.proxyUrl
            string
            URL of the HTTP proxy used for the token request.
          • spec.webhook.httpConfig.oauth2.scopes
            array of strings
            Scopes requested for the token.
            • spec.webhook.httpConfig.oauth2.scopes.Element of the array
              string
              Scope name.
          • spec.webhook.httpConfig.oauth2.tlsConfig
            object
            TLS settings for the token request.
            • spec.webhook.httpConfig.oauth2.tlsConfig.caFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.webhook.httpConfig.oauth2.tlsConfig.certFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.webhook.httpConfig.oauth2.tlsConfig.insecureSkipVerify
              boolean
              Disable verification of the server TLS certificate.
            • spec.webhook.httpConfig.oauth2.tlsConfig.keyFile
              string

              Deprecated and ignored for security reasons.

              The module does not read files from the Alertmanager container.

            • spec.webhook.httpConfig.oauth2.tlsConfig.maxVersion
              string
              Maximum TLS version.
            • spec.webhook.httpConfig.oauth2.tlsConfig.minVersion
              string
              Minimum TLS version.
            • spec.webhook.httpConfig.oauth2.tlsConfig.serverName
              string
              Server name for TLS certificate verification (SNI).
          • spec.webhook.httpConfig.oauth2.tokenUrl
            string

            Required value

            URL of the token endpoint.
        • spec.webhook.httpConfig.proxyConnectHeader
          object

          Headers of CONNECT requests to the proxy.

          Header names are mapped to lists of values.

        • spec.webhook.httpConfig.proxyFromEnvironment
          boolean
          Use the proxy parameters from the HTTP_PROXY, HTTPS_PROXY and NO_PROXY environment variables.
        • spec.webhook.httpConfig.proxyUrl
          string
          URL of the HTTP proxy for webhook requests.
        • spec.webhook.httpConfig.tlsConfig
          object
          TLS settings for requests to the webhook URL.
          • spec.webhook.httpConfig.tlsConfig.caFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.webhook.httpConfig.tlsConfig.certFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.webhook.httpConfig.tlsConfig.insecureSkipVerify
            boolean
            Disable verification of the server TLS certificate.
          • spec.webhook.httpConfig.tlsConfig.keyFile
            string

            Deprecated and ignored for security reasons.

            The module does not read files from the Alertmanager container.

          • spec.webhook.httpConfig.tlsConfig.maxVersion
            string
            Maximum TLS version.
          • spec.webhook.httpConfig.tlsConfig.minVersion
            string
            Minimum TLS version.
          • spec.webhook.httpConfig.tlsConfig.serverName
            string
            Server name for TLS certificate verification (SNI).
      • spec.webhook.url
        string

        Required value

        Webhook URL.

        The module sends notifications to this URL as JSON in the Alertmanager webhook format.

        Minimum length: 1

    • spec.zabbix
      object
      Parameters of the Zabbix channel.
      • spec.zabbix.api
        object

        Connection to the Zabbix API for automatic import of the template.

        If the section is set, the module imports the template and creates the cluster host and host group through the Zabbix JSON-RPC API. Otherwise, import the template from the ConfigMap manually.

        • spec.zabbix.api.tlsConfig
          object
          TLS settings for the HTTPS connection to the Zabbix API. Certificates and keys are given inline as PEM.
          • spec.zabbix.api.tlsConfig.ca
            string
            PEM-encoded CA certificate(s) used to verify the Zabbix API server certificate. If unset, the system trust store is used.
          • spec.zabbix.api.tlsConfig.caFile
            Deprecated
            string
            Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into ca instead.
          • spec.zabbix.api.tlsConfig.cert
            string
            PEM-encoded client certificate, for an API behind mutual TLS. Must be set together with key.
          • spec.zabbix.api.tlsConfig.certFile
            Deprecated
            string
            Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into cert instead.
          • spec.zabbix.api.tlsConfig.insecureSkipVerify
            boolean
            Disable verification of the Zabbix API server TLS certificate.
          • spec.zabbix.api.tlsConfig.key
            string
            PEM-encoded unencrypted private key for cert. Must be set together with cert.
          • spec.zabbix.api.tlsConfig.keyFile
            Deprecated
            string
            Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into key instead.
          • spec.zabbix.api.tlsConfig.serverName
            string
            Server name used for TLS certificate verification (SNI). Defaults to the host of url.
        • spec.zabbix.api.token
          string

          Required value

          Zabbix API token used to authenticate configuration.import calls.

          Minimum length: 1

        • spec.zabbix.api.url
          string

          Required value

          URL of the Zabbix JSON-RPC API, for example, https://zabbix.example.com/api_jsonrpc.php.

          If the scheme is omitted, https:// is used.

          Minimum length: 1

      • spec.zabbix.clusterName
        string

        Required value

        Name of the cluster in Zabbix.

        The name becomes the technical name of the Zabbix host that receives the data. It must be unique among all clusters and channels that send data to the same Zabbix instance.

        Minimum length: 1

      • spec.zabbix.disableTLS
        boolean

        Connect to the trapper port without TLS.

        If the parameter is disabled, the connection uses TLS with the settings from tls. If tls is omitted, the certificate is verified.

        Default: false

      • spec.zabbix.hostGroup
        string
        Zabbix host group that the cluster host is added to.

        Default: Deckhouse

      • spec.zabbix.port
        integer
        Trapper port of the Zabbix Server or Zabbix Proxy.

        Default: 10051

      • spec.zabbix.server
        string

        Required value

        Address of the Zabbix Server or Zabbix Proxy that receives trapper data.

        Minimum length: 1

      • spec.zabbix.tls
        object

        TLS settings of the connection to the Zabbix Server over the sender protocol.

        Certificates and keys are given inline as PEM. Zabbix certificate mode (TLSAccept=cert) is mutual TLS: cert and key are mandatory for it. Zabbix PSK mode is not supported.

        • spec.zabbix.tls.ca
          string
          PEM-encoded CA certificate(s) used to verify the Zabbix Server/Proxy certificate. If unset, the system trust store is used.
        • spec.zabbix.tls.caFile
          Deprecated
          string
          Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into ca instead.
        • spec.zabbix.tls.cert
          string
          PEM-encoded client certificate presented to the Zabbix Server/Proxy. Must be set together with key.
        • spec.zabbix.tls.certFile
          Deprecated
          string
          Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into cert instead.
        • spec.zabbix.tls.insecureSkipVerify
          boolean
          Disable verification of the Zabbix server TLS certificate entirely. Does not lift the client certificate requirement of Zabbix certificate mode.
        • spec.zabbix.tls.key
          string
          PEM-encoded unencrypted private key for cert. Must be set together with cert.
        • spec.zabbix.tls.keyFile
          Deprecated
          string
          Deprecated and ignored for security reasons: this filesystem path is no longer read. Put the PEM content into key instead.
        • spec.zabbix.tls.serverName
          string
          Server name used for TLS certificate verification (SNI). Defaults to spec.zabbix.server.
        • spec.zabbix.tls.skipServerNameVerification
          boolean

          Verify the server certificate chain against ca but do not match the certificate against serverName.

          Needed when the Zabbix certificate has no Subject Alternative Name (Zabbix validates only the Subject, so this is common), which is otherwise rejected. Requires ca: against the system trust store alone, a chain check without a name would accept any publicly trusted certificate. Ignored when insecureSkipVerify is true.

  • status
    object
    Status of the notification channel, including which policies are using it.
    • status.policies
      array of strings
      List of ObservabilityNotificationPolicy objects using this channel.
      • status.policies.Element of the array
        string
        Name of the notification policy.

ObservabilityNotificationPolicy

Scope: Namespaced
Version: v1alpha1

Rules for sending notifications about the alerts of one namespace.

Using the policy, the module selects alerts of the namespace where the object is created by their labels and sends notifications about them to an ObservabilityNotificationChannel from the same namespace or to a ClusterObservabilityPropagatedNotificationChannel.

  • spec
    object
    Policy parameters.
    • spec.alert
      object

      Required value

      Alerts that the policy applies to.
      • spec.alert.groupByLabels
        array of strings

        Labels by which alerts are grouped into one notification.

        Alerts with the same values of these labels are sent in one message. If the parameter is omitted, a separate notification is sent for each alert.

        • spec.alert.groupByLabels.Element of the array
          string
          Alert label name.
      • spec.alert.selector
        object

        Required value

        Selector of the alerts that the policy applies to.

        Set only one of the fields: matchLabels or matchExpressions.

        • spec.alert.selector.matchExpressions
          array of objects

          Set of expressions (conditions) for filtering alerts.

          The policy applies to an alert if the alert meets all of the conditions.

          • spec.alert.selector.matchExpressions.key
            string
            Alert label name.
          • spec.alert.selector.matchExpressions.operator
            string

            Comparison operator:

            • Equal: The label value equals value.
            • NotEqual: The label value differs from value.
            • Regex: The label value matches the value regular expression.
            • NotRegex: The label value does not match the value regular expression.

            Allowed values: Equal, NotEqual, Regex, NotRegex

          • spec.alert.selector.matchExpressions.value
            string
            Value or regular expression to compare the label value with.
        • spec.alert.selector.matchLabels
          object
          Labels and their values that an alert must have for the policy to apply to it.
    • spec.notification
      object

      Required value

      Notification parameters.
      • spec.notification.channel
        object

        Required value

        Channel that receives the notifications.
        • spec.notification.channel.kind
          string

          Required value

          Kind of the notification channel object:

          • ObservabilityNotificationChannel: A channel from the namespace of the policy.
          • ClusterObservabilityPropagatedNotificationChannel: A channel available in all namespaces.

          Default: ObservabilityNotificationChannel

          Allowed values: ClusterObservabilityPropagatedNotificationChannel, ObservabilityNotificationChannel

        • spec.notification.channel.name
          string

          Required value

          Name of the channel object.
        • spec.notification.channel.parameters
          object

          Required value

          Notification channel parameters.
      • spec.notification.repeatInterval
        string

        Required value

        Interval for resending the notification about an alert.

        Default: 1m

  • status
    object
    Notification policy state.
    • status.description
      string
      Reason why the module does not apply the policy.
    • status.state
      string

      Policy state:

      • Applicable: The module sends notifications according to the policy.
      • NotApplicable: The module does not apply the policy, for example, because the channel does not exist or parameters do not match the channel type. The reason is in the description field.

ObservabilityNotificationSilence

Scope: Namespaced
Version: v1alpha1

Temporary silence of notifications about the alerts of one namespace.

During the silence, the module does not send notifications about the alerts of the namespace where the object is created if they match spec.selector.

  • spec
    object
    Silence parameters.
    • spec.endsAt
      string

      Required value

      Time when the silence ends.
    • spec.selector
      object

      Required value

      Selector of the alerts whose notifications are silenced.

      Set only one of the fields: matchLabels or matchExpressions.

      • spec.selector.matchExpressions
        array of objects

        Set of expressions (conditions) for filtering alerts.

        Notifications about an alert are silenced if the alert meets all of the conditions.

        • spec.selector.matchExpressions.key
          string
          Alert label name.
        • spec.selector.matchExpressions.operator
          string

          Comparison operator:

          • Equal: The label value equals value.
          • NotEqual: The label value differs from value.
          • Regex: The label value matches the value regular expression.
          • NotRegex: The label value does not match the value regular expression.

          Allowed values: Equal, NotEqual, Regex, NotRegex

        • spec.selector.matchExpressions.value
          string
          Value or regular expression to compare the label value with.
      • spec.selector.matchLabels
        object
        Labels and their values that an alert must have for its notifications to be silenced.
    • spec.startsAt
      string

      Required value

      Time when the silence starts.