March 11th, 2019 2:53 PM - 4:06 PM PDT
Our ondemand and API endpoints were unreachable for our US-based product.
A subtle error in our configuration automation caused a key proxy service to stop working in our US West environment. This issue had not emerged in our other production environments, despite having deployed the same change using the same automation.
We reverted the change.
We’ve refactored the configuration automation so that this issue is no longer possible.