Instead of building, running, building, running, ... we now build all
executables once ahead of time and then run all loadtests right after
each other. This can sometimes reduce noise when load on the GHA runner
varies over time.
Since this requires us to move building into the loadtest-against
script, it also allows to go back to have the regular postgrest-loadtest
command default to building with cabal for faster local iteration.
Instead of starting a separate process, which would make the pid we
store the one from `faketime` instead of `postgrest`, we can also just
start libfaketime directly via LD_PRELOAD.
There is no need to store the out-links to the various build artifacts
in the current working directory while running postgrest-with-pg, which
is used by postgrest-loadtest. Instead reference the nix store paths
directly.
Resolves#5028
This also requires enabling shared_preload_libraries=pg_stat_statements.
Really we only need compute_query_id=on, but pg_stat_statements is
useful for other tests and it already enables compute_query_id (since
it's `auto` by default).
Instead of taking wild guesses at the runtime of the target generation
itself, we're just making sure to reset the system time to a fixed value
when we ultimately start PostgREST. This allows us to create the right
JWT expiry values ahead of time.
This was required for v9 and earlier, but these don't build with the
current nix invocation anymore anyway. Even loadtesting against v10 does
not work, because `--version` is used in one of the wait scripts and
this was only added in v11.2.
So no need to pretend we'd support comparing against older versions.
Disabling the autovacuum daemon should also help reproducibility in
theory, although I don't know of any cases where we hit a problem with
that.
VACUUM changes the order of rows that PostgreSQL returns for some table
without explicit ordering, thus doing the latter to make it consistently
reproducible.
After ANALYZE estimates are 100% exact for the moment, so some requests
which returned 206 Partial Response now return 200 instead. The fact
that PostgREST returns 206 on an unfiltered endpoint can probably be
considered a bug.
PostgREST failed when querying role settings where current
role name contained uppercase letters. This commit resolves
it by quoting the CURRENT_USER.
Signed-off-by: Taimoor Zaeem <taimoorzaeem@gmail.com>
Renames these scripts to `postgrest-with-pg-xx`. The renaming
helps reduce the length of temporary filenames. This is needed
to ensure that socket file names remain under the maximum
allowed length of 107 chars.
Closes#4461.
Signed-off-by: Taimoor Zaeem <taimoorzaeem@gmail.com>
* replaces v2-build -> v2-run (which should be v2-exec)
with one-shot v2-run
* fixes empty postgrest.prof file by sending SIGINT to the profiled
postgrest process to trigger writing it explicitly
Splitting the executable in a separate output avoids distributing all of
the libraries and documentation, which are not needed when just running
PostgREST.
Reduces closure size from 4.3G to 73.9M for the flake exported packages.
Resolves#4149
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`.
When using `postgrest-with-postgresql-* --replica`, the PGRST_DB_URI
will set the replica host as preference. This to be able to run
quick manual tests with postgrest running on a replica.
* new --replica option to `postgrest-with-postgresql-*`
* new command `postgrest-test-replica`
* new sanity tests on test_replica.py
* add postgrest-test-replica to postgrest-check and postgrest-coverage
This allows more flexible control over the working directory. Values for workingDir must always start
with a / and will then be relative to the repo root.
The upside is that postgrest-with-postgresql-xxx postgrest-test-io works as expected
now. The downside is, that postgrest-with-postgresql-xxx psql now starts without
any schema. This now needs an explicit postgrest-with-postgresql-xxx -f path/to.sql
to do anything useful.
Resolves#2864
stock curl doesn't make it easy to make many parallel requests to a
single endpoint. The endpoint has to be repeated N times.
```
curl --parallel https://example.comhttps://example.com ..
```
This provides a wrapper for parallel curl. It's useful for testing
scenarios like pool timeouts.
```
parallel-curl N https://example.com
```
By default, postgrest-with-pgrst builds postgrest as a nix package,
which means that source changes cause a full rebuild. With this
change, running the loadtest as
PGRST_BUILD_CABAL=1 postgrest-loadtest
rebuilds directly using cabal, like postgrest-build. Note that
results between nix and cabal builds aren't necessarily comparable
due to differing build parameters.
* Write vegeta output to file explicitly.
The previous version using `tee` captured the standard
output of the `withPgrst` helper, too, causing weird
errors when those had unexpected output.
* Log and fail the postgrest build explicitly.
In situations such as changing nix dependencies, the
`cabal build` run could fail, but this wasn't visible
in CI.
With these changes, the load test now fails with a nice
error message when the build fails.
Wolfgang Walther