Call me an early adopter, but I recently migrated my web service over from an Azure web-role to a "Reserved instance" on their new web sites service (mainly for git deploy!).
The migration went smoothly, but I'm seeing strange downtime patterns, typically for 5 minutes every 30 minutes (not exact timings but +/- a few minutes). According to Pingdom, the problem is DNS resolution, not the Azure service itself.
I use ZoneEdit for my DNS, have the root domain forwarded to www. www is a CNAME record which used to be linked to me.cloudapp.net (which worked fine), and is now me.azurewebsites.net
Is this DNS solution a good one? Should I be seeing this downtime? I understand that using A records directly is a bad idea due to reallocation of endpoints etc.