Best practice to fail a CD pipeline when a rollout isn't successful

[novatan-rb]

Hi everyone,

As the title suggested, I'm wondering what's the best practice to fail a CD pipeline when a rollout isn't successful.

Use case: a build goes red when the rollout is rolled back and it's in degraded state.

For kubernetes native deployments you can do it easily with kubectl rollout status, but unfortunately that doesn't work for CRDs (kubernetes/kubectl#690). I know I could whip up a little bash script to check the status in a while loop. Just wondering if a more elegant approach exists?

Thanks!

[kostis-codefresh]

Would the rollout status command work in your case? https://argoproj.github.io/argo-rollouts/generated/kubectl-argo-rollouts/kubectl-argo-rollouts_status/

[novatan-rb]

Hi @kostis-codefresh, thanks for the response. This is exactly what I need. From the look of it, the status subcommand is not available yet in current latest version (0.10.2) of kubectl-argo-rollouts?

[kostis-codefresh]

Yes you are correct. Last official release was in December 2020. I am not sure when the next version will come (I am not one of the maintainers).

[kostis-codefresh]

@novatan-rb It seems that you can get more recent binaries from here https://github.com/orgs/argoproj/packages?repo_name=argo-rollouts

[novatan-rb]

Ah cool, thanks for the help.

[jessesuen]

v1.0 is released!