Service impacts from running gitlab backup task
Everyone can contribute. Help move this issue forward while earning points, leveling up and collecting rewards.
We've recently started noticing an issue that the backup task - during at least part of the operation - is causing an outage/non-responsiveness on the https interface.
The strange thing is it's occurring for a few minutes AFTER the backup has completed (eg. backup starts at 2:00, finishes at 2:19, outage from 2:20-2:27). Almost as if the backup caused a backlog of operations, and then once it's done, that blockage is lifted, but it then floods the server handlers...
Only error I'm seeing in logs is in workhorse.log - a few of these around the time of the failures, but no other times:
2016/08/16 14:19:04 error: proxyRoundTripper: HEAD "/" failed with: "EOF"
2016/08/16 14:19:04 ErrorPage: serving predefined error page: 502I don't have any real information to support this - wondering if any of you have any ideas?