Skip to content
This repository was archived by the owner on Jun 19, 2026. It is now read-only.

receive: stagger head compaction across tenants#288

Merged
yuchen-db merged 1 commit into
db_mainfrom
yuchen-db/staggered-compaction
Feb 4, 2026
Merged

receive: stagger head compaction across tenants#288
yuchen-db merged 1 commit into
db_mainfrom
yuchen-db/staggered-compaction

Conversation

@yuchen-db

Copy link
Copy Markdown
Collaborator

This PR adds a new --receive.compaction-delay-interval flag to stagger head compaction across tenants deterministically. When set, tenant N gets a delay of N*interval (mod block duration), ensuring compactions are evenly distributed over time. This is useful for deployments with many tenants where the current random delay may still cause compaction clustering and CPU spikes. The default value of 0 preserves the existing random delay behavior.

  • I added CHANGELOG entry for this change.
  • Change is not relevant to the end user.

Changes

Verification

@yuchen-db yuchen-db merged commit e96d188 into db_main Feb 4, 2026
12 of 14 checks passed
@yuchen-db yuchen-db deleted the yuchen-db/staggered-compaction branch February 4, 2026 07:37
gavins-db pushed a commit to gavins-db/thanos-db that referenced this pull request Mar 4, 2026
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants