Remove ci_artifact_fast_removal_large_loop_limit feature flag

What does this MR do and why?

This MR removes the ci_artifact_fast_removal_large_loop_limit feature flag. We've been operating on gitlab.com with it enabled for a couple months, and haven't had any problems with throughput.

Self-hosted instances, by default, are not using the fast-removal methodology yet, and so are likely not even running into this limit. The service is still timeboxed to 5 minutes, inside a 6 minute exclusive lock, from a worker that executes every 7 minutes. Given the reactivation of expired artifact removal in %14.9 and the backports to %14.6, we're re-enabled everywhere and know that we're not running into throughput issues from backlogs on any large installations.

For context, this record-throughput limit was initially added as a specific guard against operational instability in the distributed postgres database used by gitlab.com. Self-hosted customers have not surfaced any issues that needed to be addressed with this flag.

Closes #347151 (closed)

MR acceptance checklist

This checklist encourages us to confirm any changes have been analyzed to reduce risks in quality, performance, reliability, security, and maintainability.

Edited by drew stachon

Merge request reports

Loading