The API was down for a few minutes this morning from 10:12AM to 10:20AM.

This was due to a misconfiguration in our load balancer configurations that mistakenly redirected traffic destined for api.darksky.net to the (invalid) hostname .darksky.net. This configuration had previously tested well in our testing and staging environments, but had a subtle issue in them preventing correct behavior in production.

We are adding an extra layer of tests and safeguards to catch similar issues before they impact production in the future.