usage.md

Job Creation & Docker Workflows

Assign page walkthrough

1. Select or generate a job id, choose a pool (the dropdown refreshes from worker registrations), and optionally override the timeout.
2. Add steps by selecting their StepType. When picking Docker, the form expands to capture:
   • Image: pre-built container (e.g., alpine:latest).
   • Dockerfile & Build context: for inline builds.
   • Build args: KEY=value entries passed to docker build.
   • Run args: additional flags (--network host, -p, etc.).
3. Docker steps automatically set JobPayload.Requirements.DockerRequired so only docker-capable pools accept them.
4. Submit the job and track it via the Jobs or Console pages. Each step emits [step:start]/[step:end] markers so you can trace execution and logs.

Running docker jobs locally

• The Aspire AppHost launches worker-docker with src/JD.Worker.Cli/worker-docker.yaml, sandbox mode container, and workspace isolation.
• Docker steps will:
  1. Build an image if a Dockerfile/context or build args are present. If you don't supply an image, a temporary tag (jd-worker/jobid-attempt-step) is generated and removed afterward.
  2. Mount the worker workspace at /workspace inside the container.
  3. Run the command via /bin/sh -c so multi-argument commands execute reliably even inside minimal images.

Monitoring

• Console shows workers, statuses (online/offline), and lets you pick a job to stream logs. New Logs actions fetch the latest take number of lines and allow filtering by step.
• Job detail and history pages refresh logs via Load history and Clear buttons.