Configuration

fob-worker has no config file and stores no credentials. It works on the worker repo in the current directory and follows the template's layout.

Folders#

Path	What's there
`src/steps/`	Your steps
`.orchestrator/stations/`, `.orchestrator/lines/`	Station and line files, synced with fob-orc
`.orchestrator/scenarios/<slug>/`	Saved test configs for `steps run`
`temp/`	Step outputs from local runs (`<slug>.json`) and attachment copies (`work_records/`)
`temp/stations/<STATION>/<bin>/`	Bins and workpieces

temp/ is in the template's .gitignore.

Environment variables#

The worker reads these from .env in the repo root. fob-worker loads the same .env, so a step you run locally sees the same variables as on the worker.

Variable	Required by the worker
`ORCHESTRATOR_URL`	No	The Orchestrator: `https://orchestrator.finopsbricks.com`. Defaults to `http://localhost:3000`
`ORCHESTRATOR_API_KEY`, `ORCHESTRATOR_API_SECRET`	Yes	An API key from Settings → API Keys in the Orchestrator
`WORKER_LOCATION`	Yes	Which lines this worker runs: those whose `location` matches
`POLL_INTERVAL_MS`	No	How often the worker asks for work, in milliseconds (default 2000)

Add your steps' own settings, such as API keys for the systems they call, to the same .env and read them with process.env in your step.

fob-worker itself uses none of the Orchestrator variables except WORKER_LOCATION, which steps run passes to your step as context.org_id.

Check what fob-worker sees#

fob-worker config show

Paths:
stepsDir   src/steps/
tempDir    temp

Environment Variables:
ORCHESTRATOR_URL         https://orchestrator.finopsbricks.com
WORKER_LOCATION          acme
ORCHESTRATOR_API_KEY     *** (set)
ORCHESTRATOR_API_SECRET  *** (set)

Keys and secrets are never printed, only whether they're set. Variables that aren't set are left out.

Debug output#

DEBUG=1 prints the stack trace when a command fails:

DEBUG=1 fob-worker steps run IN1_01_check --empty