Skip to content

Timeout issue during a2a agent calling #5

@olegenii

Description

@olegenii

Hi, team
It seems there is some hardcoded timeout set to 60s waiting some response from downstream agent we called and if agent executing longer - khooh emit a timeout Error.

khook installed with helm and I set 180s as api-timeout

some logs:

2025-09-17T08:57:59Z INFO event-processor Found matching hooks for event {"eventType": "helm-release-failed", "resourceName": "komoplane-20", "matchCount": 1}
2025-09-17T08:57:59Z DEBUG dedup No existing events for hook; will process {"hook": "test/helm-release-monitoring-hook", "eventType": "helm-release-failed", "resource": "komoplane-20"}
2025-09-17T08:57:59Z INFO dedup Recorded new active event {"hook": "test/helm-release-monitoring-hook", "eventType": "helm-release-failed", "resource": "komoplane-20", "firstSeen": "2025-09-17T08:57:59Z"}
2025-09-17T08:57:59Z INFO status-manager Recording event firing {"hook": "helm-release-monitoring-hook", "namespace": "test", "eventType": "helm-release-failed", "resourceName": "komoplane-20", "agentId": "kagent/aire-agent"}
2025-09-17T08:57:59Z DEBUG events Event helm-release-failed fired for resource komoplane-20, calling agent kagent/aire-agent {"type": "Normal", "object": {"kind":"Hook","namespace":"test","name":"helm-release-monitoring-hook","uid":"0f1bf798-23b3-4205-9f62-91a3ebbddbed","apiVersion":"kagent.dev/v1alpha2","resourceVersion":"209544"}, "reason": "EventFiring"}
2025-09-17T08:57:59Z INFO kagent-client Creating session for agent call {"sessionName": "hook-helm-release-failed-1758099479", "agentId": "kagent/aire-agent", "eventName": "helm-release-failed"}
2025-09-17T08:57:59Z INFO kagent-client Session created successfully {"sessionId": "ctx-76507505-b493-4a85-93a3-4a23daed44c5", "sessionName": "hook-helm-release-failed-1758099479"}
2025-09-17T08:58:15Z DEBUG workflow-coordinator Starting workflow sync
2025-09-17T08:58:15Z INFO workflow-coordinator Discovered hooks {"totalHooks": 1}
2025-09-17T08:58:15Z DEBUG workflow-coordinator No changes in hooks; keeping workflow running {"namespace": "test"}
2025-09-17T08:58:45Z DEBUG workflow-coordinator Starting workflow sync
2025-09-17T08:58:45Z INFO workflow-coordinator Discovered hooks {"totalHooks": 1}
2025-09-17T08:58:45Z DEBUG workflow-coordinator No changes in hooks; keeping workflow running {"namespace": "test"}
2025-09-17T08:58:59Z ERROR kagent-client Failed to send message to agent {"agentId": "kagent/aire-agent", "sessionId": "ctx-76507505-b493-4a85-93a3-4a23daed44c5", "error": "a2aClient.SendMessage: a2aClient.doRequest: http request failed: a2aClient.httpRequestHandler: http request failed: Post "http://kagent-controller.kagent.svc.cluster.local:8083/api/a2a/kagent/aire-agent/\": context deadline exceeded (Client.Timeout exceeded while awaiting headers)"}
github.com/antweiss/khook/internal/client.(*Client).CallAgent
/workspace/internal/client/kagent_client.go:199
github.com/antweiss/khook/internal/pipeline.(*Processor).processEventMatch
/workspace/internal/pipeline/processor.go:142
github.com/antweiss/khook/internal/pipeline.(*Processor).ProcessEvent
/workspace/internal/pipeline/processor.go:68
github.com/antweiss/khook/internal/pipeline.(*Processor).ProcessEventWorkflow
/workspace/internal/pipeline/processor.go:399
github.com/antweiss/khook/internal/workflow.(*WorkflowManager).runNamespaceWorkflow
/workspace/internal/workflow/workflow_manager.go:106
2025-09-17T08:58:59Z ERROR status-manager Recording failed agent call {"hook": "helm-release-monitoring-hook", "namespace": "test", "eventType": "helm-release-failed", "resourceName": "komoplane-20", "agentId": "kagent/aire-agent", "error": "failed to send A2A message: a2aClient.SendMessage: a2aClient.doRequest: http request failed: a2aClient.httpRequestHandler: http request failed: Post "http://kagent-controller.kagent.svc.cluster.local:8083/api/a2a/kagent/aire-agent/\": context deadline exceeded (Client.Timeout exceeded while awaiting headers)"}
github.com/antweiss/khook/internal/status.(*Manager).RecordAgentCallFailure
/workspace/internal/status/manager.go:167
github.com/antweiss/khook/internal/pipeline.(*Processor).processEventMatch
/workspace/internal/pipeline/processor.go:145
github.com/antweiss/khook/internal/pipeline.(*Processor).ProcessEvent
/workspace/internal/pipeline/processor.go:68
github.com/antweiss/khook/internal/pipeline.(*Processor).ProcessEventWorkflow
/workspace/internal/pipeline/processor.go:399
github.com/antweiss/khook/internal/workflow.(*WorkflowManager).runNamespaceWorkflow
/workspace/internal/workflow/workflow_manager.go:106

khook-config CM:

  api-timeout: 180s
  cleanup-interval-minutes: '5'
  controller_manager_config.yaml: |
    apiVersion: config.kagent.dev/v1alpha1
    kind: ControllerManagerConfig
    metadata:
      name: controller-manager-config
    health:
      healthProbeBindAddress: :8081
    metrics:
      bindAddress: :8080
    webhook:
      port: 9443
    leaderElection:
      leaderElect: true
      resourceName: khook-controller-leader-election
      resourceNamespace: kagent
    kagent:
      apiUrl: "http://kagent-controller.kagent.svc.cluster.local:8083"
      timeout: "180s"
      retryAttempts: 3
      retryBackoff: "1s"
    logging:
      level: "info"
      format: "json"
    deduplication:
      timeoutMinutes: 10
      cleanupIntervalMinutes: 5
  deduplication-timeout-minutes: '10'
  kagent-api-url: http://kagent-controller.kagent.svc.cluster.local:8083
  kagent-user-id: admin@kagent.dev
  log-level: info
  retry-attempts: '3'
  retry-backoff: 1s

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions