model-writtenrun 6052 - may be wrong, links go to the source messages. Commit-by-commit LLM analysis is in the Commits dialog.
- Koyu Tanigawa reported that requests to log the autovacuum launcher's memory contexts could take several to tens of seconds and proposed handling pending requests directly in HandleAutoVacLauncherInterrupts() #1.
- Bharath Rupireddy asked whether other processes needed similar handling and whether auxiliary processes should be supported #2; Tom Lane questioned the value of inspecting idle processes #3, while Robert Haas favored responsiveness #4 and Michael Paquier also supported responding promptly and asked about the logical replication launcher #5.
- Bharath proposed extending support to auxiliary processes with PGPROC entries #6, but Koyu explained that the logical replication launcher already used ProcessInterrupts() and requested keeping expanded process support separate from this delay fix #7.
- Fujii Masao supported handling the broader improvements separately #8, and Bharath endorsed the autovacuum patch and opened a separate thread for expanded support #9.
- Atsushi Torikoshi requested an explanatory comment matching neighboring handlers #10; Fujii added that comment, corrected indentation, and posted the revised patch with plans to commit and back-patch to PostgreSQL 14 #11.
- The revised patch #11 landed unchanged as e3e29cec10, with a PostgreSQL 14 backport; Fujii's final message confirmed the push, completing the autovacuum delay fix while broader process support remained separate #12.