2026-08-05: sidekiq_execution apdex 93.94% on urgent-cpu-bound shard
sidekiq_execution apdex 93.94% on urgent-cpu-bound shard (Severity 4)
Problem: The Sidekiq execution SLI on the urgent-cpu-bound shard is below its apdex target.
Impact: The Sidekiq execution SLI on urgent-cpu-bound dropped to 93.94% over the last 6 hours, so fewer jobs met their execution duration target on that shard.
Causes: We found that UpdateMergeRequestsWorker was driving most of the CPU usage on urgent-cpu-bound, largely because of one project with very long-running jobs.
Response strategy: We moved UpdateMergeRequestsWorker to the urgent-ci-pipeline shard, and apdex has returned to pre-incident levels without impact on that shard.
This ticket was created to track INC-12753, by incident.io