Even though the functionality "failover location" exists in WPM/Synthetic, there are situations during local disturbances, where checks can not failover to another location.
This feature request is about introducing functionality where instead of failing over checks to another location you simply run the check continuously from multiple locations and react on the combined results.
Example: You have a check for a service in UK. In the check you specify three locations you want it running from. When one location fails, you are not notified, since it's still working from 2/3 locations. Since the check result in not negative, your SLA-reports remain unaffected by the local disturbance.
Bonus functionality: Being able to look at response-time/load-time graphs and select different "perspectives" for combined multi-location checks like what does the response time look like from Europe, from UK, from US, from Manchester specifically, etc.
Real life example: Pingdom's service has this kind of functionality, at least for the graphing part.
The reacting on the combined results is a clever idea. When given time I would like to follow up on this, unfortunately I cant commit to any kind of time frame at this moment, but I will let you know when that changes.
Regarding the graphing of multiple locations for the same checks, that is easily solved by setting up the check with multiple locations and using the Analyze Metrics page.
Let me know if you have any questions!