From 21:37 through 21:40 PST, Monday July 30, our cache cluster had an issue that resulted in some requests erroring. The impacted node was replaced and service was restored.