We have completed our analysis of last Thursday’s incident that impacted some SparkPost customers provisioned in the US region.

During the incident period (20:10 UTC - 21:40 UTC), some customers experienced latency and observed an increase in 5xx error rates on calls to the Transmissions API. Additionally, some customers observed delays - up to 75 minutes in some cases - on outbound message delivery. The incident impacted only a portion of our platform tier, so many customers would not have experienced latency, errors or delay of outbound message delivery.

This incident was precipitated by an unexpected customer usage pattern that was not fully accommodated by the API service tier. We have put safeguards in place to protect against this specific scenario and we will perform a review of the entire API service tier by the end of this week to identify any usage patterns that present similar risk. We are sorry for any inconvenience this caused.