[boulos]

Sorry to hear about your troubles. Are you running with onHostMaintenance set to terminate or are you losing "regular" VMs. If you want to ping me with your project id (my username at google), I'd like to investigate. 100s of VM failures is well outside of our acceptable range.

Also, if it's been a while since your last quota request, we've drastically improved the turnaround time. All I can say is, your complaints were heard and we've tried to fix it. Keep yelling if something is busted! (And yes, I see the irony of the support ticket statement; out of curiosity which support are you on?)

Disclosure: I work on Compute Engine.