Ability to control verbosity of events

[invke]

The output of the events of the kustomize-controller quite quickly get too long on medium-sized clusters. This then causes the output in the slack provider from the notification-controller to exceed the slack message limit, truncating the event. This is a pretty easy limit to hit, I'm hitting it with ~200 resources. With any errors located at the end, the most useful information is lost from the message. This relates to an issue made on the notification-controller but probably better belongs here, as it's caused by the payload of the event sent by the kustomize-controller.

I originally thought it might be a good idea to configure the verbosity of the underlying kubectl apply that feeds into the info/error events, but that still doesn't prevent the output of the unchanged resources that floods the output.

Would it be appropriate to filter out lines that have unchanged? There are probably better solutions, like some kind of "smarter" message as mentioned by @phillebaba here, but this to me seems the simplest, albeit depending on the output of kubectl apply.

[stefanprodan]

Would it be appropriate to filter out lines that have unchanged?

I think this could work but we should filter the output only when the apply fails.