Hi, does anyone know how can I make pulumi to detect failed helm releases? Release fails but it does not detects it and doesn't try to re-release again when I re-run the pulumi

helm_chart = k8s.helm.v3.Release(
            chart_name,
            k8s.helm.v3.ReleaseArgs(
                chart=chart_path,
                version=version,
                values=helm_values,
                namespace=namespace,
                timeout=600,
                name=chart_name
            )
        )

It likely isn't noticing that it's failed on re-runs because its state isn't being refreshed. When you run

pulumi up

, are you passing

--refresh

, or do you have

options:
  refresh: always

in your

Pulumi.yaml

? If not, try it

Hi thank you for the message, I was using

pulumi refresh

before running

pulumi up.

Now I tried

pulumi up --refresh

but it did not detect the failed release again

That's odd - I'm not sure then, sorry

I couldn't fix my Pulumi problem, therefore i manually re-deployed the helm, pulumi still didn't detect any change though.

I'd recommend to use the

atomic

option for helm releases. That way, a failed upgrade gets rolled back automatically