Everytime I work with Helm in #Kubernetes I'm reminded of why I want to stay away from Helm.
I just wasted a good hour trying to figure out why our ingress-nginx helm release wouldn't deploy a service monitor. The solution was to ugrade the chart by setting `controller.metrics.enabled` to false and then true again, which triggers a new helm reconciliation.
The whole point of Kubernetes is to match a desired state with the actual state, and then Helm comes along and breaks this behavior. FML.
@garritfra That's not the worst part.
The worst part is when you try to figure out how a specific chart works and they have broken the chart down into a million and one subcharts.