Skip to content

Adds capability to specify memoryKiller per Pod

John Skarbek requested to merge jts/per-pod-memoryKiller into master
  • This allows us to be more concise when messing with memory killer settings per sidekiq deployment
  • This is important where sidekiq sharding has taken place and we'll have differing performance characteristics based on the workload

Provides the capability to address: gitlab-com/gl-infra/delivery#1084 (closed)

Merge request reports