check-workflow-health

X-McKay 的头像作者 X-McKay

编号. 检查集群中Temporal工作流的健康状况。验证Temporal服务器Pod运行状态,并检查停滞或失败的工作流。用于监控工作流编排的健康状况。关键词:Temporal、工作流、工作流停滞、工作流失败、编排。

📖 中文导读:编号. 检查集群中Temporal工作流的健康状况。验证Temporal服务器Pod运行状态,并检查停滞或失败的工作流。用于监控工作流编排的健康状况。关键词:Temporal、工作流、工作流停滞、工作流失败、编排。 以下正文为该技能的原始 SKILL.md 内容(英文),上方安装命令可直接使用。

name: check-workflow-health description: > Check health of Temporal workflows in the cluster. Verifies Temporal server pods are running and checks for stuck or failed workflows. Use for monitoring workflow orchestration health. Keywords: Temporal, workflow, workflow stuck, workflow failed, orchestration. metadata: domain: k8s category: diagnostic requires-approval: false confidence: 0.85 mcp-servers: - kubernetes-mcp-server


Check Workflow Health

Preconditions

Before applying this skill, verify:

  • Temporal is deployed in the cluster
  • Temporal namespace is known (default: temporal)

Actions

1. Check Temporal Pods

Verify Temporal server components are running.

mcp_tool: kubernetes-mcp-server/pods_list_in_namespace
params:
  namespace: temporal
timeout: 30s

2. Get Temporal Pod Logs

Check for errors in Temporal server logs.

mcp_tool: kubernetes-mcp-server/pods_log
params:
  name: temporal-server-0
  namespace: temporal
  tail: 100
timeout: 30s

3. Check Worker Pods

Verify worker pods are running and connected.

mcp_tool: kubernetes-mcp-server/pods_list_in_namespace
params:
  namespace: ai-agents
  labelSelector: app.kubernetes.io/component=worker
timeout: 30s

4. Get Temporal Events

Check for any Temporal-related events.

mcp_tool: kubernetes-mcp-server/events_list
params:
  namespace: temporal
timeout: 30s

Success Criteria

The skill succeeds when:

  • All Temporal pods are Running
  • No error patterns in recent logs
  • Worker pods are connected

Failure Handling

If Temporal is unhealthy:

  1. Check Temporal UI for workflow status
  2. Review server logs for connection issues
  3. Verify database connectivity

Examples

Output:

{
  "temporal_pods": {
    "server": "Running",
    "frontend": "Running",
    "history": "Running",
    "matching": "Running"
  },
  "workers": {
    "k8s-monitor": "Running",
    "news-monitor": "Running"
  },
  "recent_errors": [],
  "healthy": true
}