If a container gets stuck in a crash-loop-backoff, use the `logs --previous` flag
$ kubectl logs my-pod -c my-container --previous
`kubectl port-forwarding` can be useful, whether it's to check if an app is responding or to expose a remote dashboard locally (I frequently use port-forwards to look at the Argo CD dashboard!)
If you want to spin up a service locally and run and debug it like it was in the cluster, Telepresence helps here too (and can route remote traffic, volumes, and envs locally)
$ telepresence intercept my-service --port XX --mount=/tmp/
If all else fails, and you're not using distroless containers then `kubectl exec` can be useful to poke around in the container in the cluster (but remember, this is like SSHing into prod!)