@@ -110,28 +110,35 @@ For how agents respond with either a Message or a Task in A2A, see “Life of a
110110 apiVersion : kagent.dev/v1alpha2
111111 kind : Hook
112112 metadata :
113- name : pod-monitoring-hook
114- namespace : default
113+ name : basic-pod-monitoring
114+ namespace : kagent
115+ labels :
116+ environment : development
117+ monitoring-type : basic
115118 spec :
116119 eventConfigurations :
120+ # Monitor pod restarts
117121 - eventType : pod-restart
118- agentId : kagent/incident-responder
122+ agentRef :
123+ name : k8s-agent
119124 prompt : |
120- AUTONOMOUS MODE: A pod has restarted. Analyze the cause and implement fixes immediately.
121-
125+ AUTONOMOUS MODE: Pod {{.ResourceName}} restarted at {{.EventTime}}. Analyze and fix immediately.
122126 INSTRUCTIONS:
123127 • Use all available tools to diagnose and resolve the issue
128+ • Focus on common issues: crashes, resource limits, health checks, configuration
124129 • Annotate fixed resources with: kagentFix=<timestamp>
125130 • If patching fails: delete and recreate resources
126131 • Never ask for permission - assume autonomous execution
127132 • Complete remediation without human approval
133+ # Monitor OOM kills
128134 - eventType : oom-kill
129- agentId : kagent/memory-analyzer
135+ agentRef :
136+ name : k8s-agent
130137 prompt : |
131- AUTONOMOUS MODE: OOM kill detected. Analyze memory usage and optimize immediately.
132-
138+ AUTONOMOUS MODE: OOM kill for {{.ResourceName}} at {{.EventTime}}. Analyze memory and optimize immediately.
133139 INSTRUCTIONS:
134140 • Use all available tools to diagnose and resolve memory issues
141+ • Focus on memory leaks, inefficient algorithms, large allocations, GC issues
135142 • Annotate fixed resources with: kagentFix=<timestamp>
136143 • If patching fails: delete and recreate resources
137144 • Never ask for permission - assume autonomous execution
0 commit comments