v1.35.2
New Features
- Improved support for the partial readonly mode. This mode programmatically prohibits the creation, deletion, and editing of cluster-scoped objects and objects in system namespaces (
d8-*
,kube-system
). This mode is set at the build stage and is intended for the publication of the module in Community Edition and Basic Edition (CE and BE).
Bug Fixes
General
- Numerous fixes have been added for read-only access.
- Fixed terminal functionality in pods, with vim and mc now operational.
- Fixed layout extension on the YAML tab and other pages.
System
- Fixed
resourceNames
validation in cluster roles. - Fixed role creation based on an aggregated role in cluster roles.
- Fixed calculation of active roles in aggregation in cluster roles; the list of active roles is calculated recursively.
- Block Devices with references to non-existent nodes are hidden.
- Removed the ellipsis on the Kubernetes tab for Yandex next to the availability zones.
Projects
Observability
- Made the trigger threshold mandatory for completion.
- Derived metrics are better adapted to read-only rights.
General
- Only pods selected by the label selector are now shown in the service, whereas previously all pods in the namespace were shown.
- Fixed name validation in the namespace creation dialog (located in the project switcher).
Module
- Fixed
pathType
in ingress rules for Grafana.
v1.35.1
New Features
- Separate links to Commander and Stronghold have appeared on the main page if the corresponding module is enabled in the cluster
Bug Fixes
Deckhouse/modules
- Fixed module status calculation
- Added Commander source support in module overrides (ModulePullOverride)
Virtualization
- Fixed tooltip for cores and memory in the VM table
- VM snapshot breadcrumbs lead to snapshots, not disks
Projects
- Kubeconfig secrets are now displayed again in the ServiceAccount cards list
Observability
- Removed title, description, and tags from the dashboard editing form
Module
- Fixed PodDisruptionBudget parameters for frontend and grafana pods
Other
- Fixed connection hanging with clusters for Commander
v1.35.0
New Features
General
- Partial readonly mode support for Community Edition and Basic Edition.
Storage
- HPE Storage support (CSI)
- Ceph Storage support (CSI)
- Basic Kubernetes storage classes (StorageClass)
Virtualization
- VMs are displayed on cluster nodes in a separate tab
- A Monitoring tab has been added to VMs
- Machine classes — section is now in the system and in the project next to the machines themselves
- Added object source to VMs (werf, helm, etc.)
- Added a “migrate to another node” button (Evict) to VMs
- Disks can now be restored from snapshots (on the disk creation screen)
- Machine snapshots (only list and view YAML)
- Improved VM display in the general list
Deckhouse
- “Update” button on the main screen leads to the updates section
- Added placement configuration in OpenStack (currently read-only)
Observability
- Renamed “Overview” menu item to “Status”
- Trigger (alerts) and derived metrics sections added to projects
- Hide GrafanaDashboardDefinitions if the module is enabled
- The state of open and closed folders is stored in the browser
- Changes to the default data source
Modules
- Module releases are listed in the module window
Network
- Basic view of EgressGateway (list and YAML) from the cni-cilium module
- Network policies. Labels are now shown on pods/VMs to make it easier to fill in the label selector. The column with ports has been removed
Bug Fixes
General
- Fix token update (setInterval by WebWorker)
- Completed parts of the UI elements that were accessible to the user without appropriate permissions
Deckhouse
- YandexCloud settings. If the provider configuration doesn’t specify zones, all availability zones should be selected by default
- If the “operator-trivy” module is not enabled, a suggestion to enable the CVE scanner has replaced the counter on the main
Virtualization
- Improved disk
v1.34.4
Bug Fixes
Web Application
General
- Hidden user kubeconfig retrieval in Commander, will return after end-to-end rights management in Commander
- Added support for project templates (heritage=multitenancy-manager) to resource sources (checkbox)
Cluster
- Modules: Fixed documentation links for modules. Some are hidden, some lead to external sites. In the future, links will be more straightforward, but for now, they should work correctly.
Projects
- Fixed issue icons in the controller tabs “Pod:Scheduling” and “Pod:Volumes”
Observability Dashboards
- Fixed dashboard deletion
v1.34.3
Bug Fixes
Web Application
- Fixed cluster switching for Commander, now the data will be from the selected cluster
Module
- Fixed secret generation for basic-auth if HTTP mode is enabled
- Removed ingress resource with redirect from the very old name
deckhouse-admin
v1.34.2
Bug Fixes
Cluster
Deckhouse
- No longer shows an empty placeholder in the update settings when MC/deckhouse is not automatically created
Observability
- Added “Data Overview” to the menu
- “Dashboards for Grafana” section is hidden if the Observability module is enabled
- When deleting a dashboard in monitoring mode, the message “Select a dashboard” is shown
- Default mode
Mixed
anddefault_metrics
as the data source for the first panel are enabled in Data Overview - When creating a new visualization, the default data source is set to the variable
${ds_prometheus}
if it exists - Dashboard category state (opened/closed) is saved in
localStorage
for each project - GET parameters for grafana and host are synchronized in Data Overview
- Unnecessary modal windows and buttons have been removed
- Vulnerabilities have been fixed
Storage
- Sections renamed to avoid confusion with the word “volumes” in the storage class sections
General
- Web application loading now occurs on a light background, more consistent with the overall style
- Vulnerabilities have been fixed
v1.34.1
Bug Fixes
Cluster
Deckhouse
- Metrics were sometimes not shown on the main page, and the commander did not display an Overview at all
- Deckhouse update settings are not editable if the cluster is from Commander (or if there are signs of automation in MC/deckhouse at all)
- Static cluster settings are now working (Kubernetes menu item)
- In modules, the outgoing settings screen no longer knocks down the filter and sorting of lists
Virtualization
- Moved VM classes from projects to the cluster
- Images and cluster images are now created with clearer formulations.
Observability
- In grafana, when selecting a datasource, grafanaadditionaldatasources are displayed
- grafana-kube-storage has a default
timeInterval
(also known as the default$__rate_interval
) of 30 seconds, with the ability to reconfigure the environment variableSTORAGE_DEFAULT_METRICS_DATASOURCE_TIME_INTERVAL
- When selecting the menu in /explore: Add to dashboard / Existing dashboard / Select dashboard, dashboards display folder names instead of the default General
- Fixed the Not Found error that occurs periodically on open dashboards
- For filing with grafana + grafana-kube-storage, rights to read and track grafanaadditionaldatasources have been added to rbac
- The
STORAGE_DEFAULT_METRICS_DATASOURCE_TIME_INTERVAL
setting is set to{{ .Values.global.discovery.prometheusScrapeInterval }}
General
- Some items have been reduced and moved in the left navigation bar
- Fixed new sections in Commander: pages “500” and “something went wrong” won’t show up
v1.34.0
New Features
Web Application
Cluster
- The Dekhouse editorial office was added to the launch review.
- Deckhouse: Added support for the configuration of the VMware Cloud Director (VCD) provider
- Deckhouse: Added support for the configuration of the Yandex Cloud provider
- Deckhouse: Updated the deckhouse update section
- Deckhouse: Updated the global settings view, added the “Default StorageClass” field (
defaultClusterStorageClass
) - Deckhouse: Upgraded the modules section! Added update policies (MUP), Module sources (MS) and Overrides (MPO)
- Repositories — new section, will be updated further
- Storages: StorageClass classes — so far only a table and a view in YAML
- Storages: LVM-based storage management based on sds-local-volume, sds-replicated-volume, sds-node-configurator modules
- Unified the buttons
- User sessions: Instead of abstract clients, we show links to related ingresses, where possible
Projects
- In ServiceAccount, you can create kubeconfig with one click
Virtualization
- Virtual machine classes (VirtualMachineClass) — so far only a table and view in YAML
- Added the ability to place nodes in the form of a VM
- Made it easier to create images based on other images and disks
- Cluster images — a new section in administration
- A new version of badges (visual) is in the status of virtual machines - Disks and images have acquired a new modern look
Observability
- When the namespace is changed, the user remains in the same dashboard
- Fixed scroll appearance inside Grafana
- Added the “Data Overview” button in monitoring mode
- Case-insensitive search operation
- Removed PrometheusRule from projects
Server
- The API response to /cani began to include the resource that was requested. It has become more convenient to reset your license
Bug Fixes
Web Application
Cluster
- Updated links in authentication providers for OIDC and LDAP (now they are not broken, they lead to the Dex website)
- In user sessions, filtering no longer resets the selection of checkboxes
Virtualization
- Fixed bugs with access verification
- During the creation of the machine, it is clearly highlighted that a static disk is not set, so that it is obvious before saving
Server
- Fixed bugs downloading virtual images from the user’s machine
Chore
Web Application
- Updated the Content-Security-Policy header