Kubernetes Dumping Ground

Links: 111 KodeCloud Index

Events¶

We track metrics using Prometheus and Grafana.
We track application logs using ELK or fluentd.
k8s generates a lot of events related to deployment, scheduling etc.
Events are a great source of information to understand what is currently happening in our cluster.
To get the event of a particular resource we can use describe
- k describe <resource-type> <resource-name>
To get all the events we can use:
- k get events
Some important events:
- Crashloopbackoff: A container within a Pod repeatedly crashes immediately after starting up. When this happens, Kubernetes will automatically restart the container to try to recover it, but if it crashes again within a short period of time, Kubernetes will give up and mark the container as "CrashLoopBackOff".
- Imagepullbackoff: When a container within a Pod fails to pull its container image from the specified image repository. When this happens, Kubernetes will try to pull the image several times, and if it continues to fail, it will mark the container as "ImagePullBackOff".

Last updated: 2023-05-16