Hookdeck Operator Events API
Operator events are Outpost's own lifecycle and alerting stream: delivery failures, destinations being auto-disabled, retry exhaustion and tenant subscription changes. They are about the deployment, not about a tenant's traffic, so they are delivered to **operator event destinations** rather than to tenant destinations. An operator event destination is deployment-wide, is not owned by any tenant you create, and has the same shape as a tenant destination: any destination type listed by `GET /operator-events/destination-types`, with that type's `config` and `credentials`. The Operator Events API is only available on managed Outpost, and requires the Admin API Key. Self-hosted deployments configure the same thing through environment variables and a single sink. See the [Operator Events](https://hookdeck.com/docs/outpost/features/operator-events) feature page. **Available topics** | Topic | Trigger | |-------|---------| | `alert.destination.consecutive_failure` | Consecutive failure count reaches 50%, 70%, 90%, or 100% of `ALERT_CONSECUTIVE_FAILURE_COUNT` | | `alert.destination.disabled` | Destination auto-disabled at the 100% failure threshold | | `alert.attempt.exhausted_retries` | Delivery exhausts all retry attempts | | `attempt.success` | Every successful delivery attempt | | `attempt.failed` | Every failed delivery attempt, including retries | | `tenant.subscription.updated` | A destination was created, updated or deleted and the tenant's topics or destination count changed | A destination may subscribe to every topic with `["*"]`, but `attempt.success` and `attempt.failed` fire once per delivery attempt and will dominate volume. Prefer an explicit topic list. **Topics are derived from these destinations.** The union of every enabled operator event destination's `topics` is what the deployment subscribes to, which is why `PATCH /config` rejects `OPERATOR_EVENTS_TOPICS` with a 422: change the destinations and the configuration follows.