Database goals for May 14, 2018

@abrandl:

• https://gitlab.com/gitlab-org/gitlab-ce/issues/45756
  • Discussion about complexity/correctness of solution in https://gitlab.com/gitlab-org/gitlab-ce/issues/43706

Reviews:

• https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/18698

• https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/18553

• A few smaller migration reviews

• Thursday, Friday: off

@yorickpeterse:

Throughout the week:

• https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/18427 / https://gitlab.com/gitlab-org/gitlab-ce/issues/43132

Monday/Tuesday:

• https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/18427: Preload Commit authors
• Refactor Ci::Pipeline#has_warnings? so it supports eager loading of the necessary data (https://gitlab.com/gitlab-org/gitlab-ce/issues/43132#note_72170780)

Wednesday:

• Review current state of https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/18584
• Review https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/18811
• Review https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/18698
• Fix broken tests for https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/18427
• Fix pgbouncer configuration on the primary host itself.
  • It was configured to only use a pool size of 25 for web, and a size of 15 for Sidekiq.
  • Turns out that /dev/null had the wrong permissions. After fixing this we had to change the databases.json file and run gitlab-ctl reconfigure to solve the problem
  • This may have bottlenecked read-only connections going to this host since our last outage.

Thursday:

National holiday.

Friday:

• Review https://gitlab.com/gitlab-org/gitlab-ee/merge_requests/5532
• https://gitlab.com/gitlab-com/people-ops/applicant-questionnaires/merge_requests/49
• www-gitlab-com!10684 (merged)
• #77 (closed): check if vacuuming is working again later this week
• Disable loading of coverage data for https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/18427

Last week's goals: #74 (closed)