Werk #12765: k8s_pods_memory, k8s_pods_cpu, k8s_pods_fs: deprecate check plugins
|Component||Checks & agents|
|Title||k8s_pods_memory, k8s_pods_cpu, k8s_pods_fs: deprecate check plugins|
|Date||May 20, 2021|
|Checkmk Editon||Checkmk Raw (CRE)|
|Checkmk Version||2.0.0p6 2.1.0i1|
|Compatibility||Incompatible - Manual interaction might be required|
The performance problems described below only affect you, if your Kubernetes cluster exposes metrics in custom.metrics.k8s.io.
The deprecations of the plugins only affect you, if you applied the YAML files mentioned in ~/share/doc/check_mk/treasures/kubernetes/readme.md and services with the description "Memory usage for pods in NAMESPACE namespace", "CPU usage for pods in NAMESPACE namespace" or "CPU usage for Pods in NAMESPACE namespace" are present in your cluster where NAMESPACE is the name of a namespace in your cluster.
Previous versions of the Kubernetes special agent processed metrics delivered by the custom.metrics.k8s.io endpoint. The metrics were provided e.g. by applying the YAML files mentioned in ~/share/doc/check_mk/treasures/kubernetes/readme.md.
The special agent of Kubernetes could run into performance problems and timeouts if metrics were present under custom.metrics.k8s.io. Furthermore, the custom metrics were not compatible with newer versions of Kubernetes and scrape targets of the more recent special agent for Prometheus delivers data for CPU, Filesystems, Disk IO and Memory, as well. Additionally, the Prometheus agent is capable to execute custom PromQL queries for advanced use cases.
For this reason the Kubernetes special agent does not process custom metrics anymore and the accompanying check plugins and rulesets are deprecated. The corresponding services and rules should be removed. Instructions on how to use the Prometheus special agent can be found in the user guide:
https://docs.checkmk.com/latest/en/monitoring_prometheus.htmlTo the list of all Werks