What problem are you trying to solve?

Focus on the what and why of the need you have, not the how you'd like it solved.

Performance Advisor’s “Drop Index” recommendations for redundant indexes disappear after a cluster restart (or any mongod process restart). Because the recommendations aren’t persisted, there’s no reliable way to track redundant indexes over time, and they only reappear once enough workload has run again to rebuild the underlying index usage stats.

What would you like to see happen?

Describe the desired outcome or enhancement.

Redundant and unused index recommendations should survive restarts and maintenance events. Ideally, Atlas would persist the recommendations (or the underlying index usage statistics) and let users view, snooze, or dismiss them. A dedicated “Redundant Indexes” view with on-demand analysis would also help, independent of recent workload or uptime.

Why is this important to you or your team?

Explain how the request adds value or solves a business need.

We manage ~50 Atlas clusters across five regions. Maintenance, scaling, and version upgrades restart clusters regularly, so redundant index recommendations vanish before they can be reviewed or actioned. Redundant indexes add write overhead, storage cost, and memory pressure, and at our scale small inefficiencies multiply across clusters. Losing the recommendations means we can miss real optimization opportunities and can’t plan index cleanup reliably.

What steps, if any, are you taking today to manage this problem?

We export or screenshot recommendations when they appear and periodically run our own scripts (e.g., comparing listIndexes output for prefix-overlapping indexes and checking $indexStats). This is manual, per-cluster work, and $indexStats also resets on restart, so we have to time reviews carefully.