Kubernetes

인프라

클러스터를 관리하고, 워크로드를 디버깅하고, 서비스를 스케일링하세요 — Neotask이 OpenClaw으로 Kubernetes를 대화형으로 만들어줍니다.

할 수 있는 것

클러스터 상태 검사

Ask Neotask for a health summary of your cluster: which pods are not running, which nodes are under memory pressure, and whether any PersistentVolumeClaims are unbound. OpenClaw runs the right kubectl commands and presents results in plain English.

Debug Failing Workloads

Describe the symptom — "my api-server deployment keeps restarting" — and Neotask walks you through it: fetches recent events, tails the last 100 log lines, checks resource limits, and suggests the most likely fix based on what it finds.

Scale and Deploy

Tell Neotask to scale a deployment to a specific replica count, trigger a rollout of a new image tag, or pause a rollout while you investigate — no YAML editing required for routine operations.

Manage Resources Declaratively

Ask for a diff between what is running and what is in your manifests, apply a specific file, or annotate a resource — Neotask can both read and write Kubernetes resources when given permission.

Monitor Events and Alerts

Get a stream of Warning events from any namespace, see which scheduler decisions failed and why, or ask what happened to a specific pod over the past hour using describe and event history.

이렇게 물어보세요

  • "What pods are not running in the production namespace?"
  • "Why does my payment-service deployment keep crashing?"
  • "Scale the worker deployment to 10 replicas"
  • "Roll back the frontend to the previous version"
  • "Show me all OOMKilled events from today"
  • "Which nodes are above 80% memory utilization?"
  • "Drain node ip-10-0-1-42 safely before maintenance"
  • "What changed in the last rollout of the auth-service?"

    • 프로 팁

  • Specify the namespace in your prompts — "in production" or "in the staging namespace" — to avoid accidentally touching the wrong environment.
  • Ask for "a summary then details" so you get a quick answer first and can drill deeper only if needed.
  • Use Neotask to explain error messages: paste a kubectl error and ask "what does this mean and how do I fix it?"
  • Before destructive actions (drain, delete), ask Neotask to list what will be affected so you can review before confirming.
  • Combine with your CI/CD pipeline: ask "what image is currently deployed to production?" to compare against your latest build tag.

    • Works Well With