model-writtenrun 5082 - may be wrong, links go to the source messages. Commit-by-commit LLM analysis is in the Commits dialog.
- Bruce Momjian proposed documenting the autovacuum risks of pg_stat_reset() and pg_stat_reset_shared() in #1, and Robert Haas supported adding a warning in #2.
- David Rowley pointed to an earlier discussion and documentation patch in #3, noting that the warning had attracted support without objections.
- Bruce attached a reworded warning specifically about pg_stat_reset() in #4, explaining the risks of bloat and outdated statistics, but questioned whether its recovery recommendation should require VACUUM ANALYZE instead of ANALYZE.
- David demonstrated through code inspection and a SQL test that ANALYZE updates the live/dead tuple counters in #5; Bruce accepted that explanation and retained the recommendation in #6.
- Bruce committed #4 unchanged as 4d070469c1 and announced backpatching through PostgreSQL 10 in #7; David's acknowledgment in #8 closed the thread with the submitted warning landed.