Portal Community
Flow Studio

HIL Timeout — Overview

What happens when a HIL actor doesn't respond in time. Three configurable outcomes — escalate, auto-decide, or fail — ensure your workflow is never permanently blocked.

Why Timeouts Matter

Without timeouts, a HIL task waiting for a response can block a workflow indefinitely. An approver goes on leave. A form submission is forgotten. An email lands in spam. Timeouts are the safety mechanism that ensures workflows progress regardless of actor availability.

Three Timeout Outcomes

Escalate

Reassign the task to a senior actor (escalationActorId). The escalation actor gets a fresh deadline. Use for tasks that need human judgment but can be handled by a backup.

Auto-Decide

Engine injects a synthetic decision (AutoApprove or AutoReject). The workflow continues on the corresponding output port. Use when a default outcome is acceptable.

Fail

Engine throws TimeoutError — the workflow fails. Use when there is no acceptable fallback and the timeout represents a critical process failure.

Timeout Detection Architecture

ComponentResponsibility
HILTimeoutJobBackground job (Hangfire/Quartz) that scans Process_SuspendedExecutions for rows where ExpiresAt < UtcNow and Status = Pending
HILEscalationServiceResolves escalation actor, creates new HIL task, resets ExpiresAt
HILAutoDecisionServiceInjects synthetic decision, calls HILResumeService.ContinueAsync()

Timeout Job Schedule

The timeout job runs on a configurable cron schedule — default every 2 minutes. Configure in appsettings.json:

"HILTimeout": {
  "JobCronSchedule": "*/2 * * * *",   // every 2 minutes
  "BatchSize": 100                      // process up to 100 expired tasks per run
}
Precision: Timeouts are not millisecond-precise. A task with a 30-minute timeout may not be processed until up to 2 minutes after expiry (the job interval). For SLA-sensitive workflows, set the job interval to 1 minute or less.
Next: Configuring Timeout →