Managing logs at the cluster node and Pod levels

Logs are very handy when it comes to troubleshooting issues. The information collected in a log is usually helpful in understanding what has happened, figuring out why certain issues happened, and finding remediations to prevent them from happening again later on.

Cluster-level logging

In Kubernetes, the notion of cluster-level logging is widely recognized. This means logs are meant to be stored in a separate backend, so the lifecycles of those logs are independent of what’s been logged down to the worker node, pod, or even container level.

Kubernetes itself does not provide a comprehensive native logging framework, but it can be integrated with lots of third-party open source logging solutions in the community, such as Grafana Loki or the EFK stack, which includes Elasticsearch, Fluentd, and Kibana for log searching, querying, and tracing.

Logging in Kubernetes involves a set of patterns that are implemented by...