Task Timeout

Tasks support an optional timeout field that limits execution duration. When the timeout is exceeded, the controller automatically stops the Task.

Usage

apiVersion: kubeopencode.io/v1alpha1
kind: Task
metadata:
  name: fix-bug
spec:
  agentRef:
    name: my-agent
  description: "Fix the null pointer exception in auth module"
  timeout: 30m

The timeout field accepts Go duration strings: "30m", "1h", "2h30m", etc.

Behavior

• Clock starts at Running phase: The timeout is measured from status.startTime, not creation time. Queue time (Pending/Queued phases) is excluded.
• When timeout is exceeded: The controller deletes the Pod (SIGTERM), and the Task transitions to Completed with condition Stopped, reason Timeout.
• No timeout by default: If timeout is not set, the Task runs indefinitely (backward compatible).

Checking timeout status

kubectl describe task fix-bug

A timed-out Task shows:

Status:
  Phase: Completed
  Conditions:
    Type:    Stopped
    Status:  True
    Reason:  Timeout
    Message: Task timed out after 30m0s

CronTask

The timeout field works naturally with CronTask since TaskSpec is embedded in the task template:

apiVersion: kubeopencode.io/v1alpha1
kind: CronTask
metadata:
  name: daily-review
spec:
  schedule: "0 9 * * *"
  taskTemplate:
    spec:
      agentRef:
        name: code-reviewer
      description: "Review yesterday's PRs"
      timeout: 1h