Nice catch! I investigated this today with the team and we realized that metrics from a dev environment (which isn't representative of our production topology) were being included in our status page internal latency graphs for the US/EU Region Groups. We addressed the issue - you can see the clean data on status.fauna.com.