Hi Matt, Thanks for your feedback.
I really would like to unterstand the reason for Google being unable to introduce a switch labeled "route to warmed up instances only". What is so complicated with that? Personally I don't want API controls for min-idle/resident/scaling instances, as it still doesn't guarantee that requests never get routed to cold starts. As a matter of fact that's exactly what has been observed by most: idling resident instances waiting longer for doing nothing, while new dynamic instances get hit with cold starts. Further, we're using PAAS to exactly not have to take care on controlling such scaling / cold start issues. If I have to take care on it, then this would be the wrong platform for me. To me I really don't care if the instances take one, two or more seconds to boot. It's unacceptable to *ever* have requests getting directed to cold starts. Would Adwords run on this system, Google would be losing billions of dollars, if viewers would be presented with ads 2 seconds, 1 second or even few hundred milliseconds too late. We don't want that and all we're asking for is consistent performance in response times. That requires routing to warmed up instances only. Can you please spend that 1 Million it might take your engineers to adapt the scheduler to the way it should be, as it will greatly pay you off in the long run.You really have to do this, if you want to become an Amazon killer and win the crowd for your platform. Marcel -- You received this message because you are subscribed to the Google Groups "Google App Engine" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/google-appengine?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
