Precise wait-event timing and tracing
needs reviewLow-overhead core timing, trace capture, coverage analysis, independent observer-effect benchmarks, and related work to expose currently invisible waits.
postgres=# select * from postgres_hacking where visibility = 'public';
Postgres hacking,
in public.
/* agentic Postgres hacking in public — we use AI heavily, inspect every detail, test the ugly edges, and take full responsibility for the result */
5already committed upstreamAlready in Postgres and available to everyone →
[merged]transaction_timeoutStop transactions that remain open longer than the configured limit.[merged]bounded psql \watchLet psql users stop \watch automatically after a fixed number of executions.[merged]UUIDv7 generationTurn the original ULID exercise into PostgreSQL's sortable uuidv7() implementation.[merged]MultiXact standby deadlock fixFix MultiXact creation, deadlock, and corruption edge cases on standbys.[merged]COPY file and PROGRAM wait eventsExpose blocking reads and writes for COPY FROM/TO files, pipes, and programs.
latest session
archive →2026-07-15 · 50:53
Back to “log LSN on DROP TABLE”
Fix correctness and coverage of recovery landmarks for destructive DDL.
active work
view all →Log the correct commit-record LSN for DROP TABLE and related destructive operations to make PITR recovery less guessy.
Keep logical slots usable on a standby fed by WAL archives by pausing and safely resuming recovery around conflicts.
Partial materialized-view refresh
needs reviewInspired by the October 29 Hacking Postgres session, viewer Adam Brusselback implemented WHERE support for targeted REFRESH MATERIALIZED VIEW and has continued revising it through review.
Whole-WAL-record compression
blockedCompress or batch larger WAL records, originating in B-tree build experiments.
Custom protocol command tags
explorationNegotiate custom command-completion tags through a _pq_ protocol capability.
Salma El-Sayed’s implementation, mentored by Kirk Wolak, merges sparse siblings and reclaims merged-away tombstone pages without rebuilding the entire index.
pg_dump table-data placeholders
needs reviewKeep loadable slots for excluded table data so replacement data is restored before dependent indexes and constraints.
also supporting
details →TRACE / wait-event tracing
Dmitry FominDmitry’s pg_wait_tracer and core wait-event timing work, supported here with design sessions, reviews, coverage analysis, and independent benchmarks.
wait-event coverage gaps
Seongjun ShinA proposed GSoC project that continued independently: systematically identify invisible waits and add coverage, starting with blocking writes to server logging destinations.
pg_stat_statements scalability
Quan Hoang Truong + AWS contributorsQuan continued the proposed GSoC work on LWLock contention. In parallel, an AWS-led patch series implements a broader pgstat-based redesign for the same scalability problem.
recent sessions
all 45 → · 77:12 · youtube ↗
Agent hacking Postgres
Agent-assisted patch work around the wait-event timing submission.
The group reviewed Kirk Wolak's pg_dump patch for keeping loadable placeholders when table data is excluded, grounding the discussion in his large-table restore workflow. They debated a positive interface, filter-file actions, arbitrary dump queries, PII transformation, and referential-integrity risks, then posted supportive review feedback to pgsql-hackers. The final segment examined Salma El-Sayed's B-tree leaf-page merge design, including merged-page markers, scan correctness, vacuum cleanup, and review plans.
[timecodes · 8]
- Postgres 19 development opens and patches await review
- Kirk introduces the pg_dump placeholder patch
- Large-table dump and one-command restore use case
- Reviewing the option name and user interface
- Using AI to review design and mailing-list history
- Parallel chunks for restoring one large table
- Custom dump queries, filtering, and PII transformation
- B-tree leaf-page merge design and scan safety
active
correlated mapping
· 93:47 · youtube ↗
Wait-event tracing
Independent benchmarks, observer effect, coverage, and patch restructuring.
Dmitry Fomin presented a seven-patch wait-event tracing series, split for review but intended to provide complete transition timing and query attribution. The group designed an independent benchmark ladder covering vanilla, compiled-disabled, stats, and full trace modes; corrected pgbench workloads; and distinguished static overhead from observer effect. Preliminary runs showed no clear slowdown, but they added pg_stat_statements, wait-event counts, repeated rounds, raw results, and independent review before submission.
[timecodes · 8]
- Why sampled wait events miss short transitions
- Seven-patch tracing series and review structure
- Planning an independent overhead benchmark
- Designing high-transition pgbench workloads
- Five measurement levels and observer effect
- Stats versus full trace capture modes
- Correcting workload duration, rate limits, and logging
- Preliminary results and adding pg_stat_statements
active
direct mapping
· 78:11 · youtube ↗
Agentic Hacking Postgres
End-to-end tests for archive-fed logical decoding and pause/resume behavior.
The session refined logical decoding from an archive-fed physical standby as a recoverable, production-independent pipeline: after losing the decoding node, a consumer can restore another standby at the acknowledged LSN and tolerate overlap. They reviewed the three-patch implementation and TAP test, found registration and naming issues, debated lock structure, and judged the code ready for discussion after fixes. They then specified an end-to-end WAL-G test with catalog churn, paused consumption, catch-up, and a final marker.
[timecodes · 8]
- Recoverable decoding after a standby is lost
- Archive-fed WAL decoding without touching production
- Naming the feature: logical decoding from archive
- Reviewing the three-patch implementation
- TAP test registration and test readability
- Reviewing pause-on-conflict code and locking
- Patch readiness and concrete submission fixes
- Designing an end-to-end WAL-G recovery test
active
direct mapping
· 53:58 · youtube ↗
Logical replication from physical standbys on WAL shipping
Refine recovery pause behavior, crash safety, promotion, and auto-resume.
The group inspected two AI-produced patch branches for pausing archive recovery when replay would invalidate a logical slot. They reconstructed the patch series, reviewed synchronized-slot and promotion behavior, and clarified that dirty slot state is only persisted at restart points, weakening crash guarantees. The concrete conclusion was to favor correctness, consolidate the work into one reviewable three-patch series, preserve readable code over premature lock optimization, and target the July CommitFest after further testing.
[timecodes · 8]
- Returning to logical decoding on WAL-shipping standbys
- Recovering useful changes from a broken pull request
- Reviewing pause and auto-resume patches
- Conflict horizons and synchronized-slot behavior
- Dirty slot state and restart-point persistence
- Live transcript-assisted code investigation
- Restructuring the work into reviewable patches
- Crash-safety invariant and correctness-first decision
active
direct mapping