Closed ikendall closed 6 years ago
Webhook timeouts should be increased to 20 seconds since the 27th february 2018:
Our team has now deployed an increase in the webhook timeouts to 20 seconds across the board. Please let us know if you see a decrease in the error rate/sleeping webhooks after this change.
Particle team - please update the reference page of webhooks with the extended timeout: https://docs.particle.io/reference/webhooks/#error-limits
As @mguntli mentioned, we've updated all webhooks to have a 20 second timeout. With the exception of Azure IoT Hub which has a 60 second timeout.
We've observed a marked decrease in reported timeout errors since increasing the timeout.
Observed behavior
Persistent ESOCKETTIMEOUT errors in response to Webhook - caused by timeout period of 5secs. Likely cause is AWS is slower to respond than this (using Lambda). My Lambda function has to complete a number of other HTTP requests before being able to return to Webhook, which are taking longer than 5 secs.
Feature Requests
Extend the 5 secs before triggering ESOCKETTIMEOUT error to long time ... or make it configurable by the user in order to be able to match to Webhook behavior.