What to do with a broken helm release stuck applying with a high timeout?

[Weastie]

I give a high timeout to some helm releases that take a long time to rollout, for example Airflow in which the celery workers have a very high termination grace period.

What I've noticed, is that sometimes a failed helm release completely locks itself up until the timeout finishes, and I can't figure out any way to get around it. I'm sometimes just left waiting for an entire hour+.

While a failed helm release is waiting for a timeout, none of these things I tried can prematurely "fail" the release:

• kubectl delete helmrelease ...
• Removing the helmrelease from any Kustomization / yaml
• kubectl edit helmrelease and lowering the timeout to 1m

Would love to know a work around to force Flux to fail the helmrelease so it can try again.

[matheuscscp]

You are looking for the helm-controller feature gate CancelHealthCheckOnNewRevision:

https://fluxcd.io/flux/components/helm/options/#feature-gates