Kubernetes

In this catalog it is the Runtime option when you need to host and schedule agent/service workloads. Keep Kubernetes on the observe/propose side of the loop: useful signals and drafts, with human or policy authority still required before anything irreversible runs.

Kubernetes is the container runtime substrate many agent stacks deploy onto (not an agent framework).

1. Package the agent/server as container images.
2. Declare Deployments/Services/Config.
3. Schedule pods; attach volumes/secrets.
4. Scale and roll out behind the service.

Related: vllm · topics/12-central-nervous-system

• Use when: Kubernetes is the industry-standard container orchestrator: it schedules pods, heals failed instances, rolls out new versions, and exposes networking/storage primitives across a cluster matches the bottleneck and you specifically need to host and schedule agent/service workloads.
• Skip when: another pack already covers the need with less ops surface, or Kubernetes would become a second place people retype truth.
• Prefer instead: compare vLLM before locking the cell.

• Ops: confirm license, hosting, and release cadence for Kubernetes (kubernetes/kubernetes) before relying on it in a live corridor.
• Integration: bad credentials, scopes, or schema drift produce empty retrievals or runaway tool/model spend without an obvious UI alarm.
• Scope: Runtime components support agents; they are not the company ledger or the approval authority.