Sep 2026

1 commit
Fix scaling of autovacuum freeze scores.
1 Sep 2026Nathan BossartmasterunreleasedREL_19_STABLE19beta4
LLM analysis, 100% suremodified before commitlight changesfrom patchset #1442 of 2 reviewers saw the final version

Both score-greater-than-one guards in src/backend/postmaster/autovacuum.c match #144 exactly; the commit shortens the explanatory comment and rewrites the commit message. Nathan confirms the landing in #148.

  • Bharath Rupireddysaw the final versiontrailerConfirmed that the fix looked good and requested regression tests in #145, after which only explanatory text changed.
  • Greg Burdsaw the final versiontrailerChecked the arithmetic and reproduced before-and-after scores for three ages in #147, validating the logic that was committed.

Mar 2026

1 commit
Add rudimentary table prioritization to autovacuum.
27 Mar 2026Nathan Bossartmaster19beta1REL_19_STABLE19beta1
LLM analysis, 100% surelanded as postedfrom patchset #1401 of 6 reviewers saw the final version

The commit and #140 have identical added and removed lines across all seven changed files, including the all-zero-weight sorting bypass, AutoVacuumScores, GUC definitions, and documentation. Nathan announces the commit in #142.

  • Sami Imseihsaw an earlier versiontrailerTested workloads, reported the duplicate assignment, supplied documentation, and approved #132 in #133, but the explicit sorting bypass was added subsequently in #140.
  • David Rowleysaw the final versiontrailerDeveloped the threshold-ratio design, reviewed successive implementations, requested the final sorting safeguard in #139, and accepted the committed version in #141.
  • Robert Haassaw an earlier versiontrailerChallenged additive and tiered scoring in #20, requested practical testing in #60, and proposed configurable weights in #90, before the later implementation and scoring refinements.
  • wenhui qiusaw an earlier versiontrailerReported the duplicate assignment and requested score logging in #51, then discussed scaling in #108, before changes to weight limits, scaling thresholds, and component-score storage.
  • Greg Burdsaw an earlier versiontrailerSupplied simulations and an alternative comparator in #116 and approved #119 in #120, before the component-score refactoring and sorting safeguard.
  • Bharath Rupireddysaw an earlier versiontrailerReviewed documentation, observability, and testing in #126 and closely reviewed and approved #132 in #134, before the final sorting safeguard was added.

Credits across 2 commits

Author
Nathan BossartLLM2
Reviewer
Bharath Rupireddy2
Greg Burd2
David Rowley1
Robert Haas1
Sami Imseih1
wenhui qiu1
Committer
Nathan Bossart2

N LLM means N of the credits come from LLM analysis, not from commit trailers.