Avoid retries during helmreleases installations and upgrades
What does this MR do and why?
Avoid retries during helmreleases installations and upgrades
Introduce a kyverno policy that will force-reconcile failed HelmRelease once flux controller gives up.
We add several annotations to the helmrelease to ensure that force-reconciliation won't be rewuested more that 3 times.
I've been testing it with podinfo helmrelease with a bad image to see how it works:
Closes: #1358