Beginning around 13:00 PST, we noticed a spike in latency for searches and routing. After investigations, we found that this was due to AWS Lambda errors originating from the AWS service.
At around 13:21 PST Amazon reported that invoke times for the US-EAST-1 region had been resolved. We continued to monitor our data stream to ensure full recovery of systems.
At 13:36 PST we determined that the status had returned to normal and that data stream recovery had completed.
There were delays to searching and routing upwards of 15 minutes.
This was caused by AWS Lambda invoke times and errors in the US-EAST-1 region which surfaced in long delays for searching and routing within Kustomer
At 12:36 PST it was determined that Amazon had resolved their Lambda errors and systems had returned to normal
Considerations to improve upon our elastic search in the event of AWS outages.