jimmy/postgrest
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
4,361 Commits 9 Branches 102 Tags
69c6a0aa36a854a2fd2c1e31d1ecd9729598067a
Commit Graph
100 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
steve-chavez effcc080e5 chore: put ai policy first in CONTRIBUTING 
Otherwise it's not that visible.
 2026-03-23 10:10:58 -05:00
steve-chavez 622c6d3f19 fix: not logging termination unix signals 
Under container environments like ECS, it's hard to know when PostgREST
is being terminated.
 2026-03-20 13:09:10 -05:00
steve-chavezandWolfgang Walther af4c4157f5 nix: remove uneeded slocat 
slocat was introduced to test pipeline mode (see https://github.com/PostgREST/postgrest/pull/2682),
but pipeline mode was not merged https://github.com/PostgREST/postgrest/pull/2707.

So it's really not needed on the loadtests.
 2026-03-13 18:56:31 +00:00
steve-chavezandWolfgang Walther 0b7f0197fc chore: update license 2026-03-10 16:41:42 +00:00
steve-chavez 4e4ae3c7a2 chore: add AI policy in CONTRIBUTING 
To preserve our scarce review capacity, we reserve the right to reject
seemingly AI generated low-value contributions.

---

Also delete the first paragraph since now it can be interpreted as
low-effort contributions.
 2026-03-03 19:56:08 -05:00
steve-chavez 1c6d0816a5 chore: mention tests are required in CONTRIBUTING 
Also remove the Haskell conventions subtitle since these are general
recommendations.
 2026-03-02 17:57:01 -05:00
steve-chavez 66fda76644 docs: missing link to prefer:tx in config page 2026-02-19 10:23:22 -05:00
steve-chavez abe414ad88 nix: print how to tail the server logs 
For easier debugging during loadtests.

Also prefix the echo traces for the postgrest-with-pgrst command.
 2026-02-12 20:18:08 -05:00
steve-chavez e95e815483 fix: don't hide async exceptions in logs 
Fixes #4646. Using the repro on #4646, this now produces the log:

```
11/Feb/2026:09:40:08 -0500: Warp server error: stack overflow
```

When:
```
$ curl localhost:3000/
curl: (52) Empty reply from server
```
 2026-02-12 19:43:50 -05:00
steve-chavez 55cd5a23ff chore: move fixed entry to feature in changelog 2026-01-30 16:23:30 -05:00
steve-chavez ffa393850a chore: update LICENSE 2026-01-11 11:59:31 +08:00
steve-chavez 76e0e1fbff nix: fix sleep default on postgrest-with-pgrst 
The sleep was always applying since it had a default.
 2026-01-02 06:11:07 +08:00
steve-chavez b8ca1bb04c nix: add new loadtest for errors 
Also included it in CI.
 2025-12-23 15:00:24 -05:00
steve-chavez 57c5ff5db3 nix: show wait time for fixture loading 2025-12-23 15:00:24 -05:00
steve-chavez 2b83dcc794 nix: sleep param for withPgrst 2025-12-23 15:00:24 -05:00
steve-chavez 7c3bccf4ba nix: show wait time on withPgrst 2025-12-23 15:00:24 -05:00
steve-chavez d2aa6c1b8a nix: timeout param for withPgrst 2025-12-23 15:00:24 -05:00
steve-chavez 013f078bc4 fix: inaccurate Server-Timing durations 
The transaction duration was notably off, doing:

```
curl localhost:3000/rpc/sleep?seconds=5 -i
```

Shows `46.1` for the `transaction;dur`, with this fix we obtain
`5007.3`.

Fixes https://github.com/PostgREST/postgrest/issues/4522

This also fixes inaccurate "schema cache queried" logs,
see https://github.com/PostgREST/postgrest/issues/4551.
 2025-12-18 17:35:18 -05:00
steve-chavez 8f34afd66e nix: jwt loadtest chain commands and split rsa gen 
Chaining the target generation command after the build is done ensures
that if a build takes too long, the dynamic target generation for vegeta
won't cause 401 Unauthorized errors due to already expired JWTs.

Also split the rsa materials writing to another python program for
easier maintenance.
 2025-12-16 21:15:28 -05:00
steve-chavez 4cdc4c4861 nix: show build time on postgrest-with-pgrst 2025-12-16 21:15:28 -05:00
steve-chavez 050554e2bb nix: fail jwt loadtest in case of 401 
Also print last unauthorized body
 2025-12-16 21:15:28 -05:00
steve-chavez 99b8246d6e nix: add --method for postgrest-loadtest 
Currently the `postgrest-loadtest -k jwt-*` loadtests use the OPTIONS method,
which is good for isolating the perf changes for JWT auth.

But using GET is more accurate for actual perf improvements,
since it actually hits the database.

This adds a method argument for postgrest-loadtest, so GET can be used like:

```
postgrest-loadtest -k jwt-hs --method GET
```

CI is not changed, this is mostly useful for manually checking variance.
 2025-12-12 14:46:46 -05:00
steve-chavez 90bf1762ae nix: rename genTargetsHs to genTargets 
Previous name wasn't accurate as RSA is used as well.
 2025-12-12 14:46:46 -05:00
steve-chavez 018803473c bump version to 15 2025-10-24 17:39:04 -05:00
steve-chavez 23750e644f bump version to 14.0 2025-10-24 17:39:04 -05:00
steve-chavez 4dff3550b3 docs: new two-part versioning in releases section 
Also add entry to the changelog
 2025-10-24 17:23:57 -05:00
steve-chavez 3085cc81ec docs: link external auth on architecture page 
Partly addresses https://github.com/PostgREST/postgrest/issues/4110
 2025-10-24 14:56:24 -05:00
steve-chavez 4ea983163a chore: add changelog for v13.0.8 2025-10-24 14:12:02 -05:00
steve-chavez 1d2350188c chore(changelog): put additions at the top 2025-10-24 13:19:21 -05:00
steve-chavez b20d369176 docs: elaborate on Docker tag 
Also link to installation from README.md

Closes https://github.com/PostgREST/postgrest/issues/4385.
 2025-10-23 18:57:41 -05:00
steve-chavez 7968b918ec nix: fix postgrest-loadtest command 
Uses python for sending SIGNIT cross-platform
 2025-10-23 16:05:15 -05:00
steve-chavez 0f7ac1bc39 test: schema cache load max duration 
Test to ensure no regression happens after https://github.com/PostgREST/postgrest/pull/4396
 2025-10-21 14:10:14 -05:00
steve-chavezandWolfgang Walther 322216c810 chore: remove split-sections 
It's a problem for macos on x86_64 and it's no longer needed as mentioned on
https://github.com/PostgREST/postgrest/discussions/4384#discussioncomment-14614110
 2025-10-08 06:29:12 +00:00
steve-chavez 09b088d8ff change: change log-query string value to bool 
BREAKING CHANGE

As discussed on https://github.com/PostgREST/postgrest/issues/3934#issuecomment-3362806777,
this changes log-query to use a bool value, this:

- Simplifies config for users.
- Reduces effort in testing the different combinations.

Closes #3934
 2025-10-03 12:17:12 -05:00
steve-chavez e4458abd33 chore(changelog): adjust jwt sieve change 2025-10-03 12:17:12 -05:00
Steve ChavezandGitHub bcf3457c28 docs: note index on automatic tsvector (#4344) 2025-09-19 13:49:30 -05:00
Steve ChavezandGitHub 9f056ac756 fix: not logging tx variables and pre-request (#4341) 
Addresses the first item on https://github.com/PostgREST/postgrest/issues/3934
 2025-09-19 12:16:10 -05:00
steve-chavez 4a81f6b1a6 test: refactor test_log_query 
It's easier to maintain without using explicit array indexes for
matching the logs
 2025-09-17 19:08:10 -05:00
steve-chavez baea47bc85 refactor: make Query module pure 
This splits off the transaction concerns into a MainTx module

It's just moving some functions inside the Query module to the MainTx module.
Some types names were also changed so they don't have the "Query" prefix, to avoid confusion.
 2025-09-15 23:43:26 -05:00
steve-chavez 15e04903b1 fix: not logging explain query for estimated count 
Fixes https://github.com/PostgREST/postgrest/issues/4319
 2025-09-15 17:07:12 -05:00
steve-chavez a75ec75fff refactor: put CallReadPlan into CrudPlan 2025-09-15 17:07:12 -05:00
steve-chavez 0acf5a30e6 refactor: remove unnecessary RSPlan constructor 
This also removes some unreachable paths in Query module logic.
 2025-09-11 20:51:51 -05:00
steve-chavez 4650e7299f chore(changelog): add breaking change for JWT kid 2025-09-11 14:50:21 -05:00
steve-chavez 0bb127057f docs: clarify wording on kid validation 2025-09-11 14:50:21 -05:00
steve-chavez cddfb6cf5e fix: not logging OpenAPI queries 
Closes https://github.com/PostgREST/postgrest/issues/4226.

This requires moving query generation to the top App.hs module.

At this point is also simple to log the transaction variables + the
pre-request function call but this is not done here to reduce scope.
 2025-09-10 09:04:12 -05:00
steve-chavez 1d2a3e8501 refactor: pass AuthResult to txVarQuery 
Saves one parameter to the function.
 2025-09-10 09:04:12 -05:00
steve-chavez 966a611d7f refactor: inline txVarQuery and preReqQuery 
This way it's easier to see which queries are being executed.

Also use whenJust instead of traverse_.
 2025-09-10 09:04:12 -05:00
steve-chavez 1dca23e7af refactor: move accessibleFuncs to SqlFragment 2025-09-04 09:35:33 -05:00
steve-chavez 9c07070263 refactor: move accessibleTables to SqlFragment 2025-09-04 09:35:33 -05:00
steve-chavez 1ace298a1a refactor: move schemaCacheDescription to SqlFragment 
This doesn't belong in the schema cache module, as
it's unused there. It's only used on Query.hs.
 2025-09-04 09:35:33 -05:00
steve-chavez b2ce322d56 refactor: move snippet logic inside Statements.hs 
Some snippet logic was unnecessarily at the Query.hs level.

This is better as we can see what inputs we need to form the main
queries.
 2025-09-02 16:26:15 -05:00
steve-chavez 0844743c30 refactor: clarify Plan module and comment 2025-09-02 16:26:15 -05:00
steve-chavez 3e776213b7 refactor: move decoding to Query.hs 
Towards solving https://github.com/PostgREST/postgrest/issues/3934.

This makes the Statements internal module pure now it only returns
Hasql.Snippet.
 2025-09-02 16:26:15 -05:00
steve-chavez f01d5bfeb3 docs: reorder computed relationships for clarity 
Ensure the function inlining is mentioned first
 2025-08-19 12:24:47 -05:00
steve-chavez 1dce4a9321 Revert "fix: geojson invalid query on unavailable PostGIS" 
This reverts commit 0f1ca8faac.

Reverting for now as it adds one more query to the schema cache and
there's no clear way forward on how to integrate the fix with the
current schema cache queries.

See discussion on
https://github.com/PostgREST/postgrest/pull/4246#pullrequestreview-3093174224.
 2025-08-13 14:54:53 -05:00
steve-chavez 0f1ca8faac fix: geojson invalid query on unavailable PostGIS 
Closes https://github.com/PostgREST/postgrest/issues/4245.

It adds a query on schema cache construction, but doesn't add a new
attribute to the SchemaCache type. We only need the information to be
passed to the `initialMediaHandlers` function to build the builtin media
handlers.
 2025-08-06 10:20:01 -05:00
steve-chavez fa09e4af0d docs: add redirects for broken hashes on auth.rst 2025-07-30 05:28:00 -05:00
steve-chavez fd127f449c docs: correct order of jwt claims sections 2025-07-30 05:14:43 -05:00
steve-chavez ad25479185 chore: put bounded jwt cache as addition 2025-07-30 04:53:22 -05:00
steve-chavez ac155a9391 docs: clarify auth page 
Reorganizes the auth page and clarifies some links
 2025-07-28 15:43:38 -05:00
steve-chavez aa58e37f03 refactor: add PreQuery module 
Small step towards solving #3934.

Adds a module that builds SQL snippets for the transaction variables
and the pre-request function.
 2025-07-23 10:06:32 +02:00
steve-chavez 161f37f1f1 test: remove jwt decoding from mixed loadtest 
We now have dedicated JWT loadtests, so it's no longer necessary
to conflate JWT here.
 2025-07-18 07:25:53 -05:00
steve-chavez 2d4967c4b4 nix: reduce loadtest result verbosity 2025-07-16 17:38:33 -05:00
steve-chavez a183dea9e4 docs: clarify health checks empty response body 2025-07-08 12:45:15 -05:00
steve-chavez e3f8a95b72 nix: complete jwt loadtests 
Completes the jwt loadtests, now we have non-cached, cached + worst case
for the cache.

- jwt-hs (edited): now has the cache disabled
- jwt-hs-cache: cache enabled for jwt-hs
- jwt-hs-cache-worst: worst case of the jwt-hs cache
- jwt-rsa (edited): now has the cache disabled
- jwt-rsa-cache: cache enabled for jwt-rsa
- jwt-rsa-cache-worst: worst case of the jwt-rsa cache

Also deletes `nix/tools/generate_targets_rsa.py` and uses a single
python script.

Should prove what's mentioned on
https://github.com/PostgREST/postgrest/pull/4084#issuecomment-2998170423
 2025-07-06 19:14:17 -05:00
steve-chavez a87e31b767 nix: shorten loadtest script 2025-07-06 19:14:17 -05:00
steve-chavez 67bd3526c2 nix: use json file for generated jwk on loadtest 
Having the JWK as .http file was confusing.

Also split genTargets into two to avoid confusion.
 2025-06-20 11:27:44 -05:00
steve-chavez c8adfebb1f nix: fix postgrest-loadtest-to-markdown on NAs 
The python script failed when presenting NAs with:

TypeError: boolean value of NA is ambiguous
 2025-06-19 16:48:44 -05:00
steve-chavez 49de3c1ab8 nix: adjacent process monitoring report columns 2025-06-19 16:36:25 -05:00
steve-chavez e1c71c92d6 nix: add traces to loadtest commands 2025-06-19 15:04:16 -05:00
steve-chavez bc99ec8285 nix: add group to postgrest-loadtest-report 
This is so the results are easier to skim
 2025-06-17 17:58:49 -05:00
steve-chavez 7dbb8cd987 nix: better monitoring loadtest report readability 2025-06-17 16:38:06 -05:00
steve-chavez 47763df590 nix: add process monitoring for loadtests 
Closes https://github.com/PostgREST/postgrest/issues/4107.

Adds two python scripts:

- monitor_pid.py: monitors the postgrest process each second
  until it exits, then outputs a csv with the results. The nix wrappers
  use the `loadtest` dir for the output.
- merge_monitor_result.py: receives a list of csvs and merges them into
  a single markdown table. The nix wrappers use the `loadtest/*.csv`
  files for the input.

The nix `postgrest-with-pgrst` and `postgrest-loadtest-report` commands
use these scripts to add monitoring for `postgrest-loadtest` and
`postgrest-loadtest-against`.
 2025-06-17 00:10:58 -05:00
steve-chavez 53604c9db2 nix: increase jwt-hs loadtest TOTAL_TARGETS 
This is so the process monitoring results are more meaningful.

* remove the amount of total targets from the loadtest kind name
  since the targets can change.

* clarifies the top comment on the generate_targets python script

* increase TOTAL_TARGETS on jwt-hs loadtest
 2025-06-17 00:10:58 -05:00
steve-chavez ba7aaa8ab1 nix: dedup vegeta command in nix loadtest 2025-06-17 00:10:58 -05:00
steve-chavez b2f8786821 nix: detect unused python code 
Now `postgrest-lint` shows:

```
Linting workflows...
Scanning nix files for unused code...
Scanning python files for unused code...
nix/tools/generate_targets.py:13: unused variable 'JWT_DURATION' (60% confidence)
test/io/test_cli.py:6: unused import 'repeat' (90% confidence)
```

Also corrected the above detected files
 2025-06-03 20:47:39 -05:00
steve-chavez 72f6036432 docs: external authentication page in explanations 
- Move page from External JWT generation
 2025-05-30 17:59:51 -05:00
steve-chavez afc989403d docs: move jwt using SSL to tutorial 1 
Removes jwt.io example in favor of a bash script
 2025-05-30 17:59:51 -05:00
steve-chavez 054ee463a1 docs: remove Greenplum integration 
They're not really a sponsor, so it's not fair to include them.
 2025-05-28 17:43:16 -05:00
steve-chavez da467fe3b6 nix: add PGRST_ADMIN_SERVER_PORT to postgrest-run 
It's frequently needed during manual testing.
 2025-05-27 20:12:00 -05:00
steve-chavez 52cf4658c5 docs: update architecture HTTP link 
It links directly to warp page, which is confusing. Link to the
same page reference instead, which finally links to warp.
 2025-05-26 13:44:20 -05:00
steve-chavez 520631913f docs: add warning for duplicate keys in spread 2025-05-15 17:42:13 -05:00
steve-chavez 10e31c4e8c nix: clarify output of loadtest jwt 2025-05-15 16:38:59 -05:00
steve-chavez a659b0005a docs: clarify spread join table 2025-05-15 15:50:54 -05:00
steve-chavez 23f11e3abb docs: clarify fts 2025-05-15 11:57:08 -05:00
steve-chavez 211e27a4d7 docs: clarify spread feature 2025-05-15 11:14:06 -05:00
steve-chavez 257a7d1488 docs: jwt role extraction link to config 
The feature section was missing a link to its config.

Also shorten the section name.
 2025-05-09 20:45:37 -05:00
steve-chavez 98ca7c15d5 drop: Admin server config endpoint 
BREAKING CHANGE

The endpoint was at risk of being left unprotected when exposing it.

The accompanying `admin-server-config-enabled` config was also dropped.
 2025-05-04 13:53:50 -05:00
steve-chavez 7d04731be1 chore: add changelog for v12.2.12 2025-05-04 13:53:50 -05:00
steve-chavez b58253833b refactor: split preference parsing from userApiRequest 
This allows obtaining the preferences header before doing the full parse
on userApiRequest. Which is needed by #3507.
 2025-05-02 19:04:02 -05:00
steve-chavez 58237be608 test: add loadtest for async purge of JWT cache 2025-04-20 15:11:41 -05:00
steve-chavez c732591f37 chore: add changelog for 12.2.10 2025-04-18 21:30:22 -05:00
steve-chavez b740fcb1ff changelog: add missing entry jwt cache purge fix 2025-04-18 20:46:25 -05:00
steve-chavez 58b5dff188 Revert "nix: add loadtest with unique JWTs" 2025-04-18 17:43:10 -05:00
steve-chavez 51743016fd ci: github report for jwt loadtest 2025-04-18 13:05:14 -05:00
steve-chavez 9c3816218a ci: adjust upper bound of PATCH memory test 2025-04-17 23:11:33 -05:00
steve-chavez 608f7ca45a nix: add loadtest with unique JWTs 
This loadtests the jwt decoding logic. For this it adds an optional
`-k`(kind) parameter to `postgrest-loadtest` and
`postgrest-loadtest-against`.

Old kind (default):

```
postgrest-loadtest -k mixed
postgrest-loadtest-against -k mixed
```

New kind:

```
postgrest-loadtest -k jwt
postgrest-loadtest-against -k jwt
```

Internally it uses a dynamically generated targets file using python
which looks like:

```
GET http://postgrest/authors_only
Authorization: Bearer <jwt>

GET http://postgrest/authors_only
Authorization: Bearer <another-jwt>
...
```

Then this is used to run vegeta with the `-lazy` option.
 2025-04-17 23:11:33 -05:00
steve-chavez d89e7a2173 docs: redirect from broken #bulk-insert-default 2025-04-02 13:05:12 -05:00
steve-chavez 4348cb2057 docs: fix AST keyword missing from dict 2025-03-05 18:39:23 -05:00
steve-chavez 3af3371e1e docs: add note about Plan in architecture page 
Also add links to the ApiRequest, Plan and Query components in the
diagram.
 2025-03-05 18:37:10 -05:00