Atlas Real-Time Performance Panel
Parent: Database Profiler vs Query Profiler · Published reference · snapshot 2026-10-23 · skill mongodb-cohort-27f762df4c87
↓ Facts as markdownall context files
Source-anchored research for Atlas Real-Time Performance Panel. Run status: COMPLETED.
These notes link each claim to its source. Some citations identify archived research reports that are no longer hosted on this site. A published reference means the content is available; it does not certify independent review or accuracy.Read the editorial policy and follow the sources before relying on a claim.
Definitions
- RTPP is not exposed for monitoring through the Atlas Administration API, although project settings can enable or disable it through that API. (low confidence) (verified as of 2026-10-23) [source]
How it works
- Query-targeting ratios are diagnostic clues; MongoDB Search change-stream cursors can raise these ratios and trigger alerts. (low confidence) [source]
Problems, failure modes and limitations
- Use RTPP during a reproducible slow request to locate resource spikes and active operations, then investigate query details separately. A published aggregation experiment demonstrates this workflow. (medium confidence) [source]
- RTPP may fail when db.currentOp() output exceeds its 4 MB per-sample maximum. (low confidence) (verified as of 2026-10-23) [source]
- Atlas documents one-second RTPP granularity. The Administration API can enable or disable RTPP, but cannot monitor its live panel data. (low confidence) (verified as of 2026-10-14) [source]
- RTPP targets ongoing operations; Atlas separately describes Query Profiler as providing historical slow-operation samples for up to seven days. (high confidence) (verified as of 2026-10-23) [source]
Comparisons and alternatives
- COMPASS-10802 records an RTPP link returning 404 because /host/ was used instead of /metrics/. The issue was resolved with fix version 1.49.11. (low confidence) (verified as of 2026-07-01) [source]
Facts and statements
- RTPP is Atlas's live triage panel for current database operations, network traffic, and host resources; it helps correlate query latency with resource pressure. (high confidence) [source]
- RTPP requires M10+ clusters. OneUptime and a Severalnines-authored answer independently corroborate this eligibility restriction. (high confidence) (verified as of 2026-10-23) [source]
- RTPP's Slowest Operations uses db.currentOp(); Hottest Collections uses mongotop, and execution-latency charts use serverStatus opLatencies. (low confidence) [source]
- Pausing RTPP freezes its display while underlying data collection continues; resuming restarts its line graphs with an empty display. (low confidence) [source]
- The Kill Op control invokes db.killOp() on the selected running operation and requires an authorized project role. (high confidence) (verified as of 2026-10-14) [source]
- MongoDB documents one-second granularity for RTPP. (low confidence) (verified as of 2026-10-23) [source]
- RTPP uses mongostat for operation counters, mongotop for hot collections and db.currentOp() for the slowest current operations. (low confidence) [source]
- MongoDB Search change-stream cursors can inflate query-targeting ratios, so a high ratio alone does not establish an application indexing defect. (low confidence) [source]
- Queryable Encryption hides affected collections from Hottest Collections and redacts query JSON involving encrypted fields. (low confidence) (verified as of 2026-10-23) [source]
- Vokint recommends pairing Atlas live operations and cluster signals with application OpenTelemetry traces; this is an integration recommendation, not proof that RTPP provides distributed tracing. (low confidence) [source]
Children
- No children recorded.