API Reference

Incident Management API

Incidents are created automatically when a job exhausts its retries or a process stalls. Use these endpoints to list, inspect, and resolve incidents.

Overview· Workers· Incidents· Versioning· Correlation IDs· Job Policies
GET /api/v1/incidents List active incidents

Returns all active incidents across all process instances. Filter by instance, process definition, or error type.

Query parameters

ParameterTypeDescription
processInstanceKeystringFilter to a specific process instance.
processDefinitionKeystringFilter to all instances of a process definition.
errorTypestringFilter by error type: JOB_NO_RETRIES, EXTRACT_VALUE_ERROR, CONDITION_ERROR.
sizeintegerPage size, default 20, max 100.

Example response 200 OK

{
  "incidents": [
    {
      "key": "inc_aaa111",
      "errorType": "JOB_NO_RETRIES",
      "errorMessage": "Stripe timeout after 3 retries",
      "processInstanceKey": "pi_111",
      "processDefinitionKey": "payment-process",
      "jobKey": "job_xyz789",
      "elementId": "task_charge",
      "createdAt": "2026-04-06T09:05:00Z",
      "state": "ACTIVE"
    }
  ],
  "total": 1
}
GET /api/v1/incidents/{incidentKey} Get a single incident

Returns full details for a single incident by its key.

POST /api/v1/incidents/{incidentKey}/resolve Resolve an incident

Marks the incident as resolved and resumes the stalled element. Typically called after you have fixed the underlying issue (e.g. incremented job retries or corrected a variable).

No request body required. Returns 204 No Content on success.

POST /api/v1/jobs/{jobKey}/update-retries Update job retries

Sets the retry count for a job. Usually called before resolving an incident to give the job more attempts.

Request body

FieldTypeDescription
retriesintegerNew retry count (must be ≥ 1 to allow re-execution).

Example workflow

# 1. Set retries back to 3
POST /api/v1/jobs/job_xyz789/update-retries
{ "retries": 3 }

# 2. Resolve the incident to resume
POST /api/v1/incidents/inc_aaa111/resolve
Incident webhooks

You can subscribe to incident.created and incident.resolved events via the Webhooks API to trigger automated runbooks or PagerDuty alerts.

See also:   Worker API · Job policies & retry configuration · Troubleshooting guide