You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
sql/ttl: centralize row-level TTL progress updates at coordinator
Previously, each processor in a row-level TTL job directly updated job
progress, leading to fan-out issues in large clusters and performance
hits during restarts when many spans had no expired rows. Gating logic
was previously added to mitigate this.
This change delegates all progress updates to the coordinator. Each
processor now sends its progress to the coordinator, which performs the
job progress update. The original gating logic remains but is now
enforced by the coordinator.
Closes: #135227
Epic: none
Release note: none
0 commit comments