We traced the issue back to Heroku-hosted URLs. Heroku is having a router incident https://status.heroku.com/, so Hookdeck is behaving as intended. However, our Mock API is currently hosted on Heroku, so users using the Mock API will also have been impacted by those timeouts.

Only a small subset of destinations are impacted, with most timeouts originating form our Mock API (https://mock.hookdeck.com). We are still investigating the root cause.

We are investigating an issue were attempt are timing out regardless of the destination responsiveness. This seems to be caused by a networking problem within GCP. If you have a retry logic enabled, attempts will be retried. Otherwise you can bulk retry once the problem is fully resolved.