Add max retries to Prevent infinite loop in DB load balancer
Related to #344888 (closed).
-
Have a loop breaker to prevent infinite loops
When doing a DB read, we rescue connection errors and retry in while
In #344888 (closed) we had an infinite loop problem. I think we should add a max retries here say maybe a 100 max attempts ? I don't think any GitLab instance will ever have 100 hosts, or more.
-
Log when the primary host calls offline!
to help detect infinite loops !73904 (merged)
Edited by Thong Kuah