Nomad alternatives: test scheduling through failure

For Nomad alternatives, deploy a small workload with realistic placement and service-discovery needs.

1. Schedule a stateless service and a stateful job with resource limits.
2. Scale them and observe placement, health checks, and rolling updates.
3. Remove a node; verify rescheduling and data recovery.
4. Test secrets, network policy, logs, and operator visibility.
5. Compare configuration portability and cluster maintenance.

A successful initial deployment does not prove an orchestrator can survive disruption. Define recovery and isolation requirements first.