Warning

This deployment method is intended for development and testing environments only.
For production deployments, please refer to the official installation guides.

Documentation

The Bublik Docker setup documentation is available in two locations:

1. Official Documentation - Contains comprehensive setup guides and reference material

2. Instance-specific Documentation - Available at <bublik_url>/<prefix>/docs on your deployed instance

   For example: https://ts-factory.io/bublik/docs/

This documentation covers installation, configuration, usage guidelines and troubleshooting information.

E2E workflow

Four commands cover the whole loop:

Command	Does
task e2e:up	Build the images, start the E2E stack, and wait until the API, UI, logs and Celery all answer.
task e2e:seed	Generate the fixture runs and import them, skipping any the instance already has.
task e2e:test	Run the Playwright suite against it.
task e2e:down	Stop the stack, keeping its database and fixtures so the next up starts with the same data.

A typical session:

task e2e:up
task e2e:seed    # first run seeds; later runs are a no-op if the data is there
task e2e:test
task e2e:down

task e2e:up rebuilds the images, so changes in bublik-ui reach the suite only through it — the served UI is baked into the image.

Arguments after -- go to Playwright, so task e2e:test -- --grep @smoke narrows the run and task e2e:test -- --ui --ui-host=127.0.0.1 --ui-port=0 opens the interactive UI mode. task --list-all adds two more: e2e:logs (task e2e:logs -- -f celery) and e2e:types:check.

To throw everything away rather than just stopping:

docker compose -f docker-compose.yml -f docker-compose.db.yml down --volumes
python3 scripts/e2e.py clean    # fixtures, manifest, reports, traces, auth state

Both need the E2E environment — COMPOSE_PROJECT_NAME=bublik-e2e and BUBLIK_DOCKER_DATA_DIR=./data/e2e — or they will act on the production stack.

The fixture campaign

The default campaign is versioned in e2e/plan.yaml. Validate it and see what it expands to — without generating anything — with:

bublik-e2e plan --plan e2e/plan.yaml            # 39 runs, 5 dates with runs, 1 empty, ...
bublik-e2e plan --plan e2e/plan.yaml --by conclusion  # or --by fixture

Each day lists one run group per line, [fixture.]conclusion[@mix][+ui]=count:

days:
  2026-04-19: []                  # a planned empty day
  2026-04-20:
    - basic.ok@healthy=1
    - net-drv-ts.nok-warning@warn=1
    - basic.ok@healthy+ui=1       # imported through the UI form by Playwright

The plan's runs: total is asserted against what the days expand to, so an edit that adds or drops a run fails loudly. Parsing, validation and seeding all live in the bublik-e2e CLI (bublik-e2e plan/generate/run --plan e2e/plan.yaml); bublik-e2e schema --kind plan prints the plan's JSON Schema. Run the local helpers' tests with python3 -m unittest tests/test_e2e.py.

E2E uses a separate Compose project and defaults its bind-mounted data to data/e2e. The production Compose files still bind fixed host ports, so stop the normal stack before starting E2E; the two stacks cannot run concurrently without overriding all conflicting ports.