Between 10:26 AM and 02:40 PM PST, we experienced increased delays for CloudWatch log event processing for metric filter extraction and log subscriptions in the EU-WEST-1 Region. This was due to a subsystem that saw an unexpected jump in resource consumption. The issue has been resolved and the service is operating normally. New events are processing as normal, while we work through the message backlog. We expect to completely drain the backlog over the next 1 hour.