MongoDB WiredTiger Details
WiredTiger internal transactions
Data volume transfered per second between the WT cache and data files. Writes out always imply disk; Reads are often from OS filebuffer cache already in RAM, but disk if not.
Data volume handled by the WT block manager per second
Internal WT storage engine cursors and sessions currently open
A WT ‘ticket’ is assigned out for every operation running simultaneously in the WT storage engine. “Available” = hardcoded high value - “Out”.
Operations queued due to a lock.
The time spent in WT checkpoint phase. Warning: This calculation averages the cyclical event (default: 1 min) execution to a per-second value.
Least-recently used pages being evicted due to WT cache becoming full.
Configured max and current size of the WT cache.
WT internal write-ahead log operations.
Data volume moved per second in WT internal write-ahead log.
Number of records appended per second in WT internal log.
Mixed metrics: Docs per second inserted, updated, deleted or returned on any type of node (primary or secondary); + replicated write Ops/sec; + TTL deletes per second.
This panel shows the number of objects (both data (scanned_objects) and index (scanned)) as well as the number of documents that were moved to a new location due to the size of the document growing. Moved documents only apply to the MMAPv1 storage engine.
Unix or Window memory page faults. Not necessarily from mongodb.
- Page updated 2021-01-13