This is a brief post to let you know that between the 6th and 15th of June, there was an intermittent performance regression affecting our North American service area.
This was caused by a regression in our code that significantly increased database queue times when the API was under high load conditions. We were first made aware of the problem on the 8th of June by a customer but when we investigated, the high load conditions had already ceased making it difficult to replicate and diagnose the cause of the problem.
Today though we were able to view the performance regression live as it occurred and that enabled us to properly diagnose and resolve the code issue. As of this post, everything is solved and the API is once again answering all queries with a consistently low latency. We would like to thank the customers who messaged us about the problem, we would not have discovered it so quickly without your assistance :)
Thanks for reading and have a wonderful week!