Ensure we have a complete monitoring package for GitLab HA
As we introduce GitLab HA, we should ensure we have a complete monitoring story around it as well. This would include:
- Ensuring we monitor all Redis, Postgres, and nginx instances.
- Add monitoring for any extra services that are necessary like pgbouncer or repmgr.
- pgbouncer issue: #2455
- Provide any higher level metrics that would be missing, such as frequency of failovers, etc.