WORKER FLEETS
These guides cover deploying Swamp workers in container and orchestrator
environments. Each guide assumes you have a running orchestrator (swamp serve)
and want to connect workers to it.
For the initial single-worker setup, see the
Remote Execution tutorial. For the design
rationale behind fleets, see Worker Fleets.
For full CLI flag reference, see
Worker Commands. To deploy
the orchestrator on Kubernetes first, see
Deploy swamp serve on Kubernetes.
To inspect the pool visually alongside swamp worker list, open the
Workers view of the
dashboard.
In This Section
- Run a Worker in Docker — build a worker image, connect it to an orchestrator with TLS, and verify dispatch
- Run a Fleet with Docker Compose — deploy a scalable worker pool with a fleet token, configure concurrency and graceful shutdown, and run parallel workloads
- Run a Fleet on Kubernetes — deploy workers as a Deployment or Job on minikube, configure secrets and resource limits, and scale the pool
- Scale from Zero Workers — submit work to an empty pool, observe queueing, and have steps dispatch automatically when a worker connects
- Rotate a Worker Enrollment Token — keep workers connected past their token's expiry, act on the rotation report, and recover a lost token