Currently, we are recording only a few statuses in the Prometheus metrics which is incorrect. This should be done with a vector counter instead of having a metric per status.