All missed webhook events are successfully replayed to the application, restoring data consistency and ensuring no critical events like payments or updates are lost during downtime.
The agent identifies undelivered events for a specific app (e.g., Stripe) by checking the outbound status for failures like 503. It then triggers a replay of these failed events. Before acting, the agent should verify the count of undelivered events and confirm the target application slug to ensure the correct events are being recovered.
- Category: Operations and infrastructure
- Effect: Changes data in your accounts
- MCP servers: AnyHook MCP (https://koskamo.com/explore/anyhook-mcp)
- Source: AnyHook, Every webhook kept, even when your server is down (https://anyhook.net/)
- Source quote: "Our server was down last night and Stripe payments never arrived. Sort it out."
- Playbook page: https://koskamo.com/explore/use-cases/recover-undelivered-webhooks-after-server-outage
Prompt: Check for any undelivered events from the Stripe app. If there are failed deliveries, replay them to ensure they are processed.