Hi Sergey, There aren't any limits on CPU time consumed by a request, but there are limits on the wallclock time used - 30 seconds hard limit, and 1000 milliseconds for requests to be considered for scaling. Although the latter may still be in place when we switch over to the new billing, I would expect we will increase or remove it over time - it existed because billing by CPU time did not cater for requests that are long-running but use very little CPU time.
-Nick Johnson On Mon, Sep 5, 2011 at 2:48 PM, Sergey Schetinin <mal...@gmail.com> wrote: > (Another attempt to send this, because apparently the first two were > marked as spam. Trying from another email. Sorry if this is duplicate, > I waited 18hours before > sending this again.) > > After new pricing kicks in, will there still be limits on cpu usage per > request? > > I suspect the answer is "yes", so I would also like to know why other > than "that just how we wrote the thing". > > To me personally a lot of GAE limitations made much more sense in the > old pricing model because all of them (including pricing) contributed > to the impression that GAE is so efficient in stuffing machines > chock-full of instances that the hosts run busy most of the time. A > model like that requires instances to be spinned up shut down all the > time and the limits serve to make this efficient and the resource > usage more predictable. > > With the new pricing model we are told that spinning up instances is > inefficient (in fact "15min worth of resources in a couple seconds" > inefficient) and the machines are supposedly sit idle most of the time > (hence the removal of CPU charges. > > Having all of that in mind it only makes sense to remove limits on > per-request CPU usage, time limits on request handling, urlfetch > timeouts etc etc. > > Thanks. > > On 5 September 2011 07:36, Sergey Schetinin <ser...@maluke.com> wrote: > > (Second attempt to send this, because apparently the first one was > > marked as spam. Sorry if this is duplicate, I waited 18hours before > > sending this again.) > > > > After new pricing kicks in, will there still be limits on cpu usage per > request? > > > > I suspect the answer is "yes", so I would also like to know why other > > than "that just how we wrote the thing". > > > > To me personally a lot of GAE limitations made much more sense in the > > old pricing model because all of them (including pricing) contributed > > to the impression that GAE is so efficient in stuffing machines > > chock-full of instances that the hosts run busy most of the time. A > > model like that requires instances to be spinned up shut down all the > > time and the limits serve to make this efficient and the resource > > usage more predictable. > > > > With the new pricing model we are told that spinning up instances is > > inefficient (in fact "15min worth of resources in a couple seconds" > > inefficient) and the machines are supposedly sit idle most of the time > > (hence the removal of CPU charges. > > > > Having all of that in mind it only makes sense to remove limits on > > per-request CPU usage, time limits on request handling, urlfetch > > timeouts etc etc. > > > > Thanks. > > > > > > -- > http://self.maluke.com/ > > -- > You received this message because you are subscribed to the Google Groups > "Google App Engine" group. > To post to this group, send email to google-appengine@googlegroups.com. > To unsubscribe from this group, send email to > google-appengine+unsubscr...@googlegroups.com. > For more options, visit this group at > http://groups.google.com/group/google-appengine?hl=en. > > -- Nick Johnson, Developer Programs Engineer, App Engine -- You received this message because you are subscribed to the Google Groups "Google App Engine" group. To post to this group, send email to google-appengine@googlegroups.com. To unsubscribe from this group, send email to google-appengine+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/google-appengine?hl=en.
