Allow configuring batch size of UpdateAllMirrorsWorker
Everyone can contribute. Help move this issue forward while earning points, leveling up and collecting rewards.
A corrective action of https://gitlab.com/gitlab-com/gl-infra/infrastructure/issues/7173
We observed in gitlab-com/gl-infra/production#937 (closed) that when Redis is under a lot of stress, it can take UpdateAllMirrorsWorker considerable time (we observed over 100 seconds) to enqueue all ProjectImportScheduleWorker which could lead to the piling up of overdue mirrors.
Making the batch size configurable could help us mitigate such piling up by enqueuing more jobs as the enqueuing time, we also observed, doesn't change much with the number of we enqueue.
Edited by 🤖 GitLab Bot 🤖