A slow-operation threshold is an execution-time boundary used to classify database operations as slow. Atlas uses this classification to guide Performance Advisor recommendations for query optimization. section 21.2.1
Atlas provides two threshold behaviors:
- Use the Atlas-managed threshold to adjust dynamically according to operation times across the cluster. section 21.2.1
- Disable dynamic adjustment to apply a fixed 100-millisecond threshold; avoid setting this fixed threshold below 100 milliseconds. section 21.2.1
- Remember that Atlas-managed thresholds are unavailable for M0 and Flex clusters. section 21.2.1
Enable or disable dynamic classification
1. Disable Atlas-managed adjustment with `atlas performanceAdvisor slowOperationThreshold disable`. 2. Re-enable it with `atlas performanceAdvisor slowOperationThreshold enable`. 3. After identifying the relevant node, inspect its slow-query logs with `atlas performanceAdvisor slowQueryLogs list --processName <process-name>`. section 21.2.1
atlas processes list
List Atlas processes before using node-specific Performance Advisor commands. section 21.2.1
