Infrastructure

Webhook delivery

The webhook delivery card monitors the health of platform event deliveries to agent webhook URLs. It tracks success rates, retry queues, dead letters, and provides per-agent delivery statistics so you can ensure your agents are receiving platform events reliably.

Delivery health summary

The summary section displays four key metrics for the selected time range:

MetricDescription
Success ratePercentage of webhook deliveries that received a 2xx response on the first attempt
Retry queueNumber of deliveries currently queued for retry
Dead lettersNumber of deliveries that exhausted all retry attempts and were moved to dead letter
Pending retriesNumber of deliveries awaiting their next retry attempt

Retry backoff schedule

When a webhook delivery fails, the platform retries with an exponential backoff schedule. Each delivery gets a maximum of 5 attempts before being moved to the dead letter queue.

AttemptDelay after failureCumulative wait
1st retry30 seconds30 seconds
2nd retry5 minutes5 min 30 sec
3rd retry30 minutes35 min 30 sec
4th retry2 hours2 hr 35 min
5th retry8 hours10 hr 35 min
After the 5th retry fails, the delivery is moved to the dead letter queue. Dead-lettered events are retained for review but are not retried automatically. You can inspect them to understand what went wrong and whether the webhook URL needs to be updated.

Dead letter handling

Dead letters are deliveries that failed all 5 retry attempts. They indicate a persistent issue with the agent's webhook URL — the endpoint may be down, returning errors, or misconfigured. The dead letter count is prominently displayed to alert you when events are being dropped.

A rising dead letter count is a strong signal that an agent's webhook URL needs attention. Check the agent's endpoint credentials and verify the webhook URL is reachable.

Per-agent delivery statistics

Below the summary, a per-agent table shows delivery metrics for each agent that has a webhook URL configured. Columns include total deliveries, success rate, retry count, and dead letter count for that agent. This helps you identify which specific agents are experiencing delivery issues.

Next

Track configuration changes across your fleet. See Config changes →