The module is intended for the basic monitoring of cluster nodes.
It safely collects metrics and provides a basic set of rules for monitoring of:
- The current Docker version on the node (and if it complies with the requirements);
- The overall health of the cluster monitoring subsystem (Dead man’s switch);
- The availability of file descriptors, sockets, abundance of free space and inodes;
- The operation of
- The state of cluster nodes (NotReady, drain, cordon);
- The state of time synchronization between nodes;
- The cases of the prolonged CPU stealing;
- The state of the Conntrack table on nodes;
- The Pods that report an incorrect state (due to kubelet-related or other issues), etc.