Backport of https://gitlab.com/gitlab-org/gitlab-ee/merge_requests/14597

See merge request gitlab-org/gitlab-ce!30520

This is the first part of Docker Registry replication
    for secondary Geo node.

This will help diagnose the source of excessive I/O from Rugged
calls. To implement this, we need to obtain the full list of arguments
sent to each request method.

Remove dead mysql code

Closes #63191

See merge request gitlab-org/gitlab-ce!29608

None of this code can be reached any more, so it can all be removed

This is no longer needed with Rails 5.2. opclass is the attribute used
per https://github.com/rails/rails/pull/19090/files.

Now that we've removed the monkey patch and restored the Rails schema
dumper, it appears Rails has dropped the inclusion of `using: :btree` as
well (https://github.com/rails/rails/pull/27981).

Closes https://gitlab.com/gitlab-org/gitlab-ce/issues/64529

Add Rugged calls and duration to API and Rails logs

Closes #64676

See merge request gitlab-org/gitlab-ce!30871

Use relative worker identifier for metrics (instead of Process.pid) and
identify when Unicorn/Puma/Sidekiq is used.
Previously, it was assumed that all metrics are gathered from Unicorn
due to hardcoded implementation which was incorrect.

This adds `rugged_duration_ms` and `rugged_calls` fields to
`api_json.log` and `production_json.log`. This will make it easier to
identify performance issues caused by excessive I/O.

Closes https://gitlab.com/gitlab-org/gitlab-ce/issues/64676