Change reactive caching urgency to high
What does this MR do and why?
Change reactive caching urgency from low to high to improve merge request widget performance.
Customer Impact
Users are experiencing significant delays when loading merge request widgets (license, security, and testing reports), with response times exceeding 30 seconds in some cases. The root cause is that ReactiveCachingWorker has low urgency, causing it to be deprioritized in the Sidekiq queue behind other work. Low urgency workers have a queue scheduling target of 1 minute, which is too long for work that is powering UI elements that users are actively waiting on. High urgency workers have a target of 10 seconds and should be scheduled even faster when the queue is short. This should improve the responsiveness of the MR widgets significantly.
Technical Justification
Analysis of the worker's performance over the last 7 days shows:
- Median job execution time: consistently less than 0.5 seconds
- 99th percentile execution time is usually under 10 seconds (with occasional spikes)
These metrics meet the requirements for high urgency workers as defined in the Sidekiq worker attributes guidelines:
- Median execution time < 1 second ✓
- 99% of jobs complete within 10 seconds ✓
Performance graph showing execution time distribution:
References
Related to https://gitlab.com/gitlab-com/request-for-help/-/issues/3890
MR acceptance checklist
Evaluate this MR against the MR acceptance checklist. It helps you analyze changes to reduce risks in quality, performance, reliability, security, and maintainability.