jingrow/gunicorn
1
0
Fork 0
mirror of https://github.com/frappe/gunicorn.git synced 2026-07-01 18:21:30 +08:00
Code Issues Packages Projects Releases Wiki Activity
gunicorn/tests/docker/per_app_allocation/README.md
Benoit Chesneau 1d0df29796 feat(dirty): add class attribute workers support and e2e tests 
- Add get_app_workers_attribute() to read workers class attribute
- Update _parse_app_specs() to check class attribute when no config override
- Add Docker-based e2e tests for per-app worker allocation
- Add test apps: HeavyModelApp (workers=2), LightweightApp
- Add unit tests for get_app_workers_attribute function
- Add integration tests for class attribute detection
2026-02-01 03:04:35 +01:00

1.6 KiB
Raw Blame History

Per-App Worker Allocation E2E Tests

End-to-end Docker-based tests for the per-app worker allocation feature.

Overview

These tests verify that:

  • Apps with worker limits are only loaded on the specified number of workers
  • Requests are routed only to workers that have the target app loaded
  • Round-robin distribution works correctly within limited worker sets
  • Worker crash scenarios maintain correct app allocation
  • Class attribute workers=N is respected
  • Config-based :N overrides class attributes

Configuration

The tests use 4 dirty workers with 3 apps:

  • LightweightApp: No limit (loads on all 4 workers)
  • HeavyApp: workers=2 class attribute (loads on 2 workers)
  • ConfigLimitedApp: :1 config (loads on 1 worker)

Running Tests

# From this directory
cd tests/docker/per_app_allocation

# Build the Docker image
docker compose build

# Run all tests
pytest test_per_app_e2e.py -v

# Run specific test
pytest test_per_app_e2e.py::TestPerAppAllocation::test_config_limited_app_uses_one_worker -v

Test Categories

TestPerAppAllocation

  • Tests basic functionality of per-app worker allocation
  • Verifies round-robin distribution
  • Tests app accessibility

TestPerAppWorkerCrash

  • Tests behavior when workers crash
  • Verifies app recovery after worker respawn

TestPerAppLogs

  • Verifies logging output contains expected information

Requirements

  • Docker and Docker Compose
  • Python 3.8+
  • pytest
  • requests

Notes

  • Tests run on port 8001 to avoid conflicts with the existing dirty_arbiter tests on 8000
  • The container uses a keep-alive wrapper to allow testing worker crash scenarios