Receive Prometheus Alertmanager events

Accepts the standard version-4 Alertmanager webhook with at most 20 alerts and no truncation. Requires both incidents:ingest and incidents:write on the current API key; test/live isolation follows the key. Keep sourceId stable for each independent Alertmanager source and shared across its HA replicas. No Idempotency-Key is required. Per-alert fingerprint and nanosecond startsAt ordering deduplicate retries and fence delayed recovery. A recovery received before firing records a terminal source episode without creating an incident. Repeated firing cannot reopen a manually resolved episode; a newer startsAt may open another incident. The whole batch, its clock, incident cancellation and queue intent commit atomically. The body is limited to 256 KiB. See /docs/incident-alertmanager for receiver configuration.

Requestcurl -X POST https://api.transmit.dev/v1/integrations/alertmanager/{configurationId}/{escalationPolicyId}/{sourceId} \
  -H "Authorization: Bearer tr_v2_…" \
  -H "Content-Type: application/json" \
  -d '{
       "version": "4",
       "truncatedAlerts": 0,
       "alerts": [
         {
           "status": "firing",
           "fingerprint": "0011223344556677",
           "startsAt": "2026-10-21T00:00:00Z",
           "endsAt": "0001-01-01T00:00:00Z",
           "labels": {
             "alertname": "DatabaseDown",
             "severity": "critical"
           },
           "annotations": {
             "summary": "Database unavailable"
           }
         }
       ]
     }'

Parameters

Body

Responses

Every response body is described in the OpenAPI spec, which this page is generated from.