First you need to know why the POD
was evicted. You can do that by kubectl describe pod <pod name>
.
From there you will get the eviction reason, which can be checked with Eviction Signals
If you are running on just one node, it's possible you run out of resources (space/cpu/ram).
After that you should do kubectl describe node <node-name>
that will show what type of resource cap the node is hitting under Conditions:
section.
If that won't helm and POD
will still get evicted you should check logs bykubectl logs <pod-name> -f
and see if there are anything helpful.