Postmortem -
Read details
May 25, 22:25 UTC
Resolved -
We have already rolled back this configuration change and all data is being processed again as normal as of 1:53pm ET. We'll be sharing a post-mortem shortly.
May 24, 17:53 UTC
Investigating -
At 1:31 PM ET (5.31pm UTC) we started to roll out an infrastructure upgrade to improve our ability to scale the Knock API. This upgrade caused an unexpected failure in API processing for 22 minutes until 1:53 PM ET due to a misconfiguration around a critical component of our API ingestion layer.
During this time frame, workflow triggers, and integration events were not processed and were dropped, despite successful API responses. Scheduled workflows, changes made to Knock environments using the CLI or dashboard, and already stored user data were not affected. Any affected workflow triggers or source events sent during this window can be resent to our API to ensure they are processed.
May 24, 17:31 UTC