Issue Summary

In this task, the test logs include a large number of WT_VERB_CACHE_TOP INFO messages such as:

sweep-server: [WT_VERB_CACHE_TOP][INFO]: cache top dirty leaf bytes: 0 tables above 4096B hold 0B of 0B in use, 3221225472B configured
sweep-server: [WT_VERB_CACHE_TOP][INFO]: cache top recent bytes evicted: 0 tables above 4096B hold 0B

Context

• The logs were introduced as part of cache top diagnostic work related to WT-18566 and WT-18400.
• A comment in WT-18566 already mentions this verbosity concern.
• current behavior is:
  • verbose level 1 prints only the top 5 tables per pass
  • verbose level 2 prints all 32 tracked tables
• The concern is that the test should not print all verbose cache_top logs by default, or the amount of output at verbose level 1 should be reduced.

Proposed Solution

Reduce cache_top log noise in default test output. Possible fixes:

• Do not print all verbose logs in tests by default.
• Reduce the amount of log lines emitted at verbose level 1 for cache_top.

Definition of done

Either justification for why logging doesn't need to be reduced, or the logging is reduced for test format.

References

• Related Jira issues: WT-18566, WT-18400
• Slack discussion: Slack thread

Original Slack thread: https://mongodb.slack.com/archives/C0ANJ0L2QEM/p1793060964277769
This ticket was generated by AI from a Slack thread.