If you edit the console operator config and enable a random plugin name (or the name of a real plugin that doesn't have a ConsolePlugin resource for whatever reason), the console will crash loop. This seems to be a regression introduced by the new internal pluginsOrder argument to the console backend. The console operator will skip plugins without ConsolePlugin resource, but still adds the plugin to pluginsOrder.

For example, this error:

E0820 18:14:49.393104 1 sync_v400.go:726] failed to get "pipelines-console-plugin" plugin: consoleplugin.console.openshift.io "pipelines-console-plugin" not found

Results in the following console-config config map values:

{{
plugins: []
pluginsOrder:

• pipelines-console-plugin
  }}

And the following error from the console pod:

F0820 17:40:43.219455 1 validate.go:76] invalid flag: plugins-order, error: list must only contain currently enabled plugins