Internal Website Checks - Change alerting behavior when one collector does down

It looks like if a collector goes down, the internal website check won't trigger the alert for website-overall, if we configure alert trigger to "at All test locations" and have more than one collector assigned to monitor the app. 

Strangely, I can see in the raw data and graphs that the website-Overall does switch value to 2 (i.e. not OK). 

We are an AWS shop and deploy collectors in each client's VPC AND from a collector that has routing to the client VPC (e.g. from the client's on-prem network). We were conducting DR testing where the AWS based collector was "killed" along with the rest of the VPC, but we never saw an alert trigger for website-Overall.