Fix WriteAheadLog DB health check for placeholder user record reassignment

Everyone can contribute. Help move this issue forward while earning points, leveling up and collecting rewards.

• Work on this issue
• Close this issue

With :reassignment_throttling feature flag enabled, there have been errors in the WriteAheadLog DB health status indicator

The error has only occurred in our reassignment service (exported log):

We use this indicator to throttle placeholder user record reassignment service if/when necessary. Initial investigation indicates we are not passing the connection argument to the execute method correctly, or perhaps not setting the context up properly.

From Slack (90 days):

The WriteAheadLog calls execute on the connection, which is derived from the context https://gitlab.com/gitlab-org/gitlab/-/blob/df310bc767a0e84b7042af056b2b6e81a2850a8e/lib/gitlab/database/health_status/indicators/write_ahead_log.rb#L63