So far, when enabling automatic service monitor deployment, the CI build/test caused remotely by a PR to operatorhub.io tend to require prometheus (which should be OK), but then things usually (but not always) fail. The current (temporary?) solution is to require the user to hand install the service monitor from susql-operator/hack/susql-servicemonitor.yaml when installing from operatorhub. It isn't a huge burden because it seems that the servicemonitor only needs to be installed once, but it is not elegant to require the user to take that extra step....
This should be super easy, but it isn't.
So far, when enabling automatic service monitor deployment, the CI build/test caused remotely by a PR to operatorhub.io tend to require prometheus (which should be OK), but then things usually (but not always) fail. The current (temporary?) solution is to require the user to hand install the service monitor from
susql-operator/hack/susql-servicemonitor.yamlwhen installing from operatorhub. It isn't a huge burden because it seems that the servicemonitor only needs to be installed once, but it is not elegant to require the user to take that extra step....