Add a cluster-level boot smoke test for the computing-unit-worker

### Task Summary

`amber-integration` boot-smoke-tests `texera-web` (#6319) and `computing-unit-master` (#6377), but the **computing-unit-worker** is not covered — and it can't use the same `smoke-boot.sh` port-probe approach:

- `ComputingUnitWorker.main` only calls `AmberRuntime.startActorWorker`, which binds `pekko.remote.artery.canonical.port = 0` (random) and starts no HTTP server — so there is no fixed port to probe (unlike the master's Dropwizard `:8085`).
- A worker's health is "joined the cluster," which requires a seed/master at `:2552` first. Standalone it just retries and stays up, so a "didn't crash" check proves nothing.

A meaningful worker smoke test needs to bring up a master + worker and assert the worker actually joins the cluster (cluster membership), not merely that the process survives. Track adding that.

Raised from #6377 review.

### Task Type

- [ ] Refactor / Cleanup
- [x] DevOps / Deployment / CI
- [x] Testing / QA
- [ ] Documentation
- [ ] Performance
- [ ] Other