awesome-postgres
A community catalog of PostgreSQL software, libraries, tools and resources, kept by Asad Dhamani since 2015-08-02 and published both as a GitHub repository and as the rendered page this ingest read. 12,052★, 1,003 forks, CC0-1.0, 67 open issues, last pushed 2026-05-21. The description names its ancestor: "inspired by awesome-mysql."
It is the spoke's third tooling source and its second broad directory, after awesome-db-tools —
which is the condition synthesis.md said would reopen the tooling-coherence question. It does, and
this time the catalog earns its place on a different ground than coverage.
The shape of the list is the finding
Roughly thirty sections and several hundred entries, and the distribution across them is lopsided in a way worth writing down. Counted from the rendered page:
| What the section holds | Sections | Entries (approx.) |
|---|---|---|
| Alternative servers — forks and rebuilds of the engine | Server | 5 |
| Extensions running inside the engine | Extensions | ~34 |
| Tools around a stock server — GUI, CLI, backup, HA, monitoring, optimization, utilities, queues, K8s | 10 | ~220 |
| Managed services | PaaS, Docker, Distributions | ~24 |
| Client bindings | Language Bindings | 17 languages |
| Reading — tutorials, blogs, podcasts, newsletters, docs, community | 9 | ~55 |
Five entries in the entire catalog are a different server: AgensGraph (graph), Apache Cloudberry (an MPP fork, positioned as a Greenplum alternative), FerretDB (a MongoDB-compatible front end on Postgres), Postgres-XL, and YugabyteDB (distributed SQL on a PostgreSQL fork). Everything else either runs inside Postgres or beside it.
That ratio is the ecosystem's argument about itself, and it is the same argument citus made from one vendor's side. Citus is a distributed database that refused to leave Postgres because the extension seam was wide enough; this catalog is the population-level version of that claim — when people want Postgres to do something new, they overwhelmingly write an extension or a sidecar rather than a fork. Written up as postgresql-extensions.
What it is good for, and what it is not
T3. It is an index, not an analysis: one line per entry, no dates, no maintenance signal, no
opinion about which of forty GUI clients to use. Its value here is the same as awesome-db-tools' —
a source of page candidates — and this ingest used it that way twice over.
Its weakness is the one every link list has, and it is visible in the text. The whole catalog
carries exactly one staleness marker: wal-e is annotated "obsolete" and its successor wal-g
listed separately. Nothing else says whether an entry shipped last week or was abandoned in 2017, and
several sections (GUI, Utilities) mix commercial products, dormant projects and current tools in one
alphabetical run. The list was last pushed three months before this ingest. So every claim in the
eleven deep pages below comes from the project's own repository or documentation, never from the
catalog line — and every claim in the fourteen coverage pages is explicitly marked as the catalog's
word, unverified.
How this catalog was ingested
Two layers, deliberately.
Eleven subjects read at their primary source (T1), chosen because each closes a gap the spoke had already named rather than because it was the most popular entry:
| Page | Section it came from | What it closed |
|---|---|---|
| patroni | High-Availability | The failover path; where consensus actually lives |
| pgbackrest | Backups | The write path (growth edge 5) |
| pgbouncer | Utilities | A claim postgresql was asserting unsourced |
| pgmq | Work Queues | Postgres-instead-of-a-broker |
| yugabytedb | Server | The commit protocol (growth edge 1); Raft in a database |
| ferretdb | Server | The document neighbour (growth edge 8), partly |
| hypopg | Optimization | What-if index testing |
| pg-stat-statements | Monitoring (substrate) | Which query is expensive, cumulatively |
| pgbench | Monitoring (misfiled) | The measurement standard the spoke kept lacking |
| write-ahead-logging | — (engine docs) | WAL, needed by two of the above |
| postgresql-planner-statistics | — (engine docs) | Growth edge 9: where the estimate comes from |
| logical-replication | — (engine docs) | The other half of replication |
And every remaining entry recorded in fourteen coverage pages, section by section, at the catalog's word, so nothing in the list is unaccounted for: postgres-ha-tooling · postgres-backup-tooling · postgres-monitoring-tooling · postgres-query-optimization-tooling · postgres-gui-clients · postgres-cli-tooling · postgres-schema-migration · postgres-work-queues · postgres-data-movement · postgres-client-libraries · postgres-managed-services · postgres-kubernetes-operators · postgres-forks-and-derivatives · postgres-learning-resources. The Extensions section's full roster lives on postgresql-extensions.
The split is the honest one: what was verified is marked T1 and cited to a primary; what was only catalogued is marked T3 and says so on the page.
Cross-spoke context
Two sections sit largely outside this spoke and are recorded rather than claimed:
- PaaS (18 entries) — the managed-hosting and pricing angle belongs to
../cloud-wiki; roster in postgres-managed-services, which keeps only the architectural half (what a managed platform does to the extension model). The PlanetScale for Postgres line is a fact about a vendor this spoke pages as a MySQL sharding house (planetscale). - Kubernetes operators (9) — running a database as a Kubernetes workload is
../platform-ops-wiki's practice; roster in postgres-kubernetes-operators, which keeps the architectural half (they are mostly packaging patroni).
Related
postgresql · postgresql-extensions · patroni · pgbouncer · pgbackrest · pgmq · yugabytedb · ferretdb · hypopg · pg-stat-statements · pgbench · write-ahead-logging · postgresql-planner-statistics · logical-replication · awesome-db-tools · pev2 · citus · operational-databases · synthesis