You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
podListInstruction=fmt.Sprintf("- Use pods_list_in_namespace with namespace parameter set to '%s' to get all pods in namespace '%s'", namespace, namespace)
43
44
}
44
45
45
46
verboseMsg:=""
@@ -77,7 +78,7 @@ Follow these steps systematically:
77
78
- Count total nodes and categorize issues
78
79
79
80
## 3. Check Pod Health
80
-
- Use pods_list to get all pods%s
81
+
%s
81
82
- Identify problematic pods:
82
83
* Phase = Failed or Pending (CRITICAL)
83
84
* Container state waiting with reason:
@@ -161,7 +162,7 @@ Warnings: [count]
161
162
- Be efficient: don't call the same tool multiple times unnecessarily
162
163
- If a resource type doesn't exist (e.g., ClusterOperator on vanilla K8s), skip it gracefully
163
164
- Provide clear, actionable insights in your summary
164
-
- Use emojis for visual clarity: ✅ (healthy), ⚠️ (warning), ❌ (critical)`, scopeMsg, namespaceFilter, verboseMsg)
165
+
- Use emojis for visual clarity: ✅ (healthy), ⚠️ (warning), ❌ (critical)`, scopeMsg, podListInstruction, verboseMsg)
165
166
166
167
assistantMessage:=`I'll perform a comprehensive cluster health check following the systematic approach outlined. Let me start by gathering information about the cluster components.`
0 commit comments