kubectl --dry-run=server

[eyarz]

any specific reason why you are covering --dry-run=client and not --dry-run=server?
--dry-run=server will produce better results when trying to validate the Kubernetes files:
https://github.com/datreeio/kubernetes-schema-validation#misconfigurations-coverage-summary

[HighwayofLife]

dry-run=client does not need a kube config to run, whereas dry-run=server does. I do agree that server gives much better validation and results than any other validation tool, so to add support for dry-run=server, we need to add support for a kube-config/context.

[eyarz]

dry-run=client also requires a running K8s cluster to work...

[HighwayofLife]

Related issue. kubernetes#51475. --dry-run=client should not require access to the kube API.

[eyarz]

Interesting!
I didn't know that.