Launching webhooks
For Day 4 of our launch week, we're excited to announce that webhooks are out of beta and are now generally available! You can try them out on Hatchet Cloud or in your self-hosted deployment. You can read the docs here.
Overview
Webhooks are simple to build initially but can be surprisingly difficult to run at scale. Unexpected failures, high latency, and dropped requests can all lead to a poor user experience or missing data. As a result, webhook requests are often placed into a queue so that they can be retried and replayed at some point in the future.
We discovered that teams were using our Events feature to have webhooks kick off Hatchet tasks, so that webhooks could be retried, monitored and replayed directly from Hatchet. To make things easier, we decided to build a webhook proxy layer in front of Hatchet.
Local Development
One of the things we're particularly excited about is that you no longer need a local tunnel like ngrok to test webhook infrastructure. Your worker already has a persistent connection to the Hatchet engine, so the only thing you need to do is configure the webhook in Hatchet. This also happens to be more secure, as the worker connection is client-initiated and uses TLS to connect via gRPC.
Getting Started
You can read the docs on how to set up webhooks here, along with cookbooks for specific integrations:
We'd love to hear your feedback on this feature, so please report any issues in the GitHub repo or reach out to us on Discord if you have any questions or run into any problems!