Action item

Investigate and remediate the recurring server-side causes of majority replication lag seen in repeat-offender clusters.

Context

Provides the engineering disposition for customers that remain affected after the Stream 1 prevention rollout — that is, clusters where the released pacing and prevention mechanisms are active and majority replication lag still recurs.

Ownership

• Stream: Stream 4 — Create a repeat-offender process with CSM and TSE support
• Team(s): Replication
• Owner: Unassigned — to be agreed at rollout kickoff
• Rollout gate: Phase 0 — Mobilize (runs throughout)

Background

This is a tracked action item for the Atlas roll-out of majority replication lag prevention, detection, and remediation — Milestone 3 of the Majority Replication Lag SMART goal (GOAL-251), tracked by SPM-4794.

Milestone 1 established fleet-wide observability and RCA foundations. Milestone 2 delivered server enhancements and lag-aware pacing mechanisms. Milestone 3 turns those results into an operational system that detects incidents, explains their causes, gives COEs an approved response path, and reduces recurrence among the worst offenders.

References

• Tracking epic: SPM-4794
• SMART goal: GOAL-251
• Rollout plan: GOAL-251: Atlas Rollout (section 2.5)
• Repl Lag SMART Goal Plan: Milestone plan
• Related: SERVER-129638
• Related: SPM-4631, SPM-4519 (prevention work whose residual gaps this ticket absorbs)
• Related: Engineering Proposal: Capping Majority Replication Lag for 8.0+ Clusters

Filed from a placeholder ticket reference in the rollout plan document. Title, scope, and acceptance criteria are expected to be refined by the owning team.