Batch large pipelined Redis commands
What does this MR do and why?
This MR breaks large pipelines in Gitlab::Redis::Cache
into batches of 1000.
It uses the same idea as reactive_set_cache's keys.each_slice(1000) { |subset| pipeline.unlink(*subset) }
. In Redis, pipelines sizes should be capped since the server holds the responses in memory and huge pipelines could place an unexpectedly large load on the server as seen in the incident issue below.
See https://gitlab.com/gitlab-com/gl-infra/reliability/-/issues/23699. This is CA for gitlab-com/gl-infra/production#14406 (closed).
Does not affect self-managed as it is behind a feature-flag.
Screenshots or screen recordings
Screenshots are required for UI changes, and strongly recommended for all other merge requests.
Before | After |
---|---|
How to set up and validate locally
Numbered steps to set up and validate the change are strongly suggested.
MR acceptance checklist
This checklist encourages us to confirm any changes have been analyzed to reduce risks in quality, performance, reliability, security, and maintainability.
-
I have evaluated the MR acceptance checklist for this MR.