一覧へ戻る
NikolayS/postgres_dba（新しいタブで開きます）
主要言語: PLpgSQL
The missing set of useful tools for Postgres DBAs and all engineers
- star 数
- 1,293
- watcher 数
- 46
- fork 数
- 144
- issue 数
- 6
- 最終更新
- 2026/08/19
gem-hunter主要言語: PLpgSQL
The missing set of useful tools for Postgres DBAs and all engineers
34 diagnostic reports for PostgreSQL, right inside psql. No agents, no daemons, no external dependencies — just SQL.
Bloat estimation, index health, lock trees, vacuum monitoring, query analysis, corruption checks, buffer cache inspection, and more. Type :dba and go.
git clone https://github.com/NikolayS/postgres_dba.git cd postgres_dba bash -c "echo \"\\set dba '\\\\\\\\i $(pwd)/start.psql'\" >> ~/.psqlrc"
Connect to any Postgres server via psql and type :dba.
Requires psql 10+. The server can be any version. For best results, use the latest psql client.
| ID | Report |
|---|---|
| 0 | Node info: primary/replica, replication lag, database size, temp files, WAL, replication slots |
| 1 | Database sizes and stats |
| 2 | Table and index sizes, row counts |
| 3 | Load profile |
| ID | Report |
|---|---|
| a1 | Current connections grouped by database, user, state |
| l1 | Lock trees (lightweight) |
| l2 | Lock trees with wait times (PG14+ pg_locks.waitstart) |
| ID | Report |
|---|---|
| b1 | Table bloat estimation |
| b2 | B-tree index bloat estimation |
| b3 | Table bloat via pgstattuple (expensive) |
| b4 | B-tree index bloat via pgstattuple (expensive) |
| b5 | Tables without stats (bloat can't be estimated) |
| ID | Lock | Report |
|---|---|---|
| c1 | AccessShareLock | Quick index check: btree + GIN (PG18+). Safe for production. |
| c2 | AccessShareLock | Indexes + heap/TOAST (PG14+). Safe but reads all data. |
| c3 | ⚠️ ShareLock | B-tree parent check — detects glibc/collation corruption. Use on clones. |
| c4 | ⚠️⚠️ ShareLock | Full: heapallindexed + parent + heap. Proves every tuple is indexed. |
| ID | Report |
|---|---|
| m1 | Buffer cache contents (pg_buffercache, expensive) |
| ID | Report |
|---|---|
| i1 | Unused and rarely used indexes |
| i2 | Redundant indexes |
| i3 | Foreign keys with missing indexes |
| i4 | Invalid indexes |
| i5 | Index cleanup DDL generator (DO & UNDO) |
| ID | Report |
|---|---|
| v1 | Vacuum: current activity |
| v2 | Autovacuum progress and queue |
| ID | Report |
|---|---|
| p1 | CREATE INDEX / REINDEX progress |
| ID | Report |
|---|---|
| s1 | Slowest queries by total time |
| s2 | Full query performance report |
| s3 | Workload profile by query type |
| ID | Report |
|---|---|
| t1 | Postgres parameters tuning |
| t2 | Objects with custom storage parameters |
| e1 | Installed extensions |
| x1 | Alignment padding analysis (experimental) |
| r1 | Create user with random password |
| r2 | Alter user with random password |
Some reports benefit from additional extensions:
| Extension | Reports | Install |
|---|---|---|
pg_stat_statements |
s1, s2, s3 | shared_preload_libraries = 'pg_stat_statements' |
amcheck |
c1, c2, c3, c4 | CREATE EXTENSION amcheck; |
pgstattuple |
b3, b4 | CREATE EXTENSION pgstattuple; |
pg_buffercache |
m1 | CREATE EXTENSION pg_buffercache; |
Tested on PostgreSQL 13 through 18 via CI on every commit. Older versions (9.6–12) may work but are not actively tested.
Works with the pg_monitor role — superuser is not required for most reports (corruption checks need superuser or explicit GRANT EXECUTE).
Drop a .sql file in sql/. The filename format is <id>_<name>.sql. The first line must be a -- comment with the description — it becomes the menu entry automatically.
# Regenerate the menu after adding/removing reports
bash ./init/generate.shpspg makes tabular output much easier to read:
\setenv PAGER pspg
\pset border 2
\pset linestyle unicodeBuilt on diagnostic queries contributed by many people over the years:
Nikolay Samokhvalov — nik@postgres.ai