Takashi: Like Richard, I really don't understand; perhaps we are missing something obvious; if so, I apologize in advance. The scenario that Jeff laid out is exactly what I would expect, but from what I can gather, that is NOT how App Engine is currently working.
For my particular application right now it is not a huge problem, but there are many scenarios where it might become one. More importantly however, it just doesn't make sense to me, and that is a big warning sign. What else might change in a way that makes no sense? On Sunday, July 15, 2012 9:04:10 PM UTC-4, Takashi Matsuo (Google) wrote: > > > I would say, the best bet is to set an appropriate number of min idle > instances. > > Tom, > > > The resident instance handles almost no requests (only 3 so far). If a > > dynamic instance needs to be started, couldn't the scheduler send at > > least some requests to the idle instance while calling /_ah/warmup on > > the spawning dynamic one? > > No at lease for now, because the scheduler can not tell beforehand how > long the warmup request will take, and even whether the warmup request will > succeed or not, thus those subsequent requests will see the significant, > un-predictable pending time and probably elevated error rate in some cases. > > Michael, > > > Under load, I wouldI want user facing requests to hit the resident > instance UNTIL a new dynamic instance has been spun up AND hit with a > warmup request. > > If you set enough number of min idle instances to absorb your trafic, it > should be Ok. > > In my understanding, technically, it might be possible for us to provide > prediction based warmup request, but it will cause complaints from many > people who want to save money. Now we're offering configurable min idle > instance, so please consider using this feature. > > If you have any use-case or experience which can not be covered by setting > min-idle-instances, please let us know. > > -- Takashi > > > On Mon, Jul 16, 2012 at 12:18 AM, Simon Knott wrote: > >> Completely agree, seems to defeat the entire purpose of warm-up requests. >> >> >> On Sunday, 15 July 2012 14:40:14 UTC+1, Michael Hermus wrote: >>> >>> I have to agree with this; it seems completely backwards to me. Wouldn't >>> resident instance warmups be extremely infrequent since they are.... well, >>> resident! Under load, I wouldI want user facing requests to hit the >>> resident instance UNTIL a new dynamic instance has been spun up AND hit >>> with a warmup request. >>> >>> >>> -- >> You received this message because you are subscribed to the Google Groups >> "Google App Engine" group. >> To view this discussion on the web visit >> https://groups.google.com/d/msg/google-appengine/-/kV3oMH_pUloJ. >> >> 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. >> > > > > -- > Takashi Matsuo > -- You received this message because you are subscribed to the Google Groups "Google App Engine" group. To view this discussion on the web visit https://groups.google.com/d/msg/google-appengine/-/3D2T4XvWl-YJ. 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.
