Hi Per, Thanks for sending the details. First thing comes to mind is that the culprit might be instances on a version 'ah-builtin-python-bundle'. This version is currently dedicated to the 'Datastore Admin' feature. I think you are running some task like backup/copy/deletion in 'Datastore Admin', right?
I hope it could explain the situation you are experiencing right now. Please let me know if that isn't the case. Thanks, -- Takashi On Tue, May 15, 2012 at 7:06 AM, Per <per.fragem...@gmail.com> wrote: > <https://lh3.googleusercontent.com/-qa-cjA-cUbM/T7GBSShwGrI/AAAAAAAAADY/OYljQYgcczc/s1600/changed-2-weeks-ago.jpg> > > > Hi Takashi, > > I waited with my response because I didn't want to jump to conclusions. > Yes, I have been experimenting with some settings every now and then, > limiting instances, playing with the pending sliders etc. But all it did > was change how quickly instances were created or collected (while always > increasing latency). It didn't change how the billed instances were > calculated. > > > Above is the chart of my applications' instances for the past month. (The > editor didn't allow me to place it at the right location...) It's rather > easy to spot when things changed. We used to get billed for roughly one > instance on average. Yes, we did have more than one running, but it was > (and is!) always one instances that handles 95% to 99% of the load. So it > seemed only just that Google would only charge for the main instances. I > never asked for an instance that just sits there, but I didn't mind it > while it was free :) > > > So now being charged for two or three instances, when only one is really > doing anything, seems like a major change that should be documented. Okay, > maybe it's just our application, but our pricing has increased steeply. I > just posted another message about the frontend hour calculation. Both > issues combined seem to have led to a pricing increase from $5 to $10 > before, to now $30 to $50. Our request per second have increased > moderately, we may have made some requests slower, we might have slightly > different usage patterns. But I cannot see a reason for a price increase > this steep. > > Any help or insight would be appreciated. Our app ID is > small-improvements-hrd > > Kind regards, > Per > > > > > On Friday, April 27, 2012 7:02:26 AM UTC+2, Takashi Matsuo (Google) wrote: >> >> On Fri, Apr 27, 2012 at 10:16 AM, Per <per.fragem...@gmail.com> wrote: >> > Hi team, >> > >> > previously, the scheduler used to spin up 3 instances for our >> > application, only used 1 out of the 3, but at least we didn't have to >> > pay for the unused instances. There was always a big gap between total >> > instances and billed instances. >> >> Wasn't it because that you set Max Idle Instances at that time? >> If you set Max Idle Instances to Automatic, the total instances and >> billed instances should be the same. >> >> > >> > However, as of yesterday, the scheduler continues to spin up 3 >> > instances on average, but we have to pay for all of them. We're at >> > maybe 1 request per second on average, and it's all handled just fine >> > by one instance. We're on F4, so having to pay for 2 mostly unused >> > instances hurts. >> > >> > >> > Since this seems to coincide with the 1.6.5 release, I'm tempted to >> > think that there's something changed in the background, and it would >> > be great to learn more about the new suggested way. >> > >> > When we tried to limit idle instances last time, all we got was >> > instance churn and bad latency. So we're on automatic/automatic these >> > days, and I'd prefer not to have to experiment again. Some advice >> > would be great. >> >> Again, I'm wondering when you tried to limit the idle instances. >> Generally speaking, you can not get everything. If you want a better >> performance, you'll need to pay more, and vice versa. >> >> However, if you're really certain that the behavior of our scheduler >> has significantly changed, please let me know your app-id and detailed >> explanation(it can be off-list), so I can look into it further. >> >> > >> > That aside 1.6.5 is great of course! :) >> >> Thanks! >> >> -- Takashi >> >> > >> > Cheers, >> > Per >> > >> > -- >> > 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 <google-appengine@googlegroups.com>. >> > To unsubscribe from this group, send email to >> google-appengine+unsubscribe@**googlegroups.com<google-appengine%2bunsubscr...@googlegroups.com>. >> >> > For more options, visit this group at http://groups.google.com/** >> group/google-appengine?hl=en<http://groups.google.com/group/google-appengine?hl=en>. >> >> > >> >> >> >> -- >> Takashi Matsuo | Developer Advocate | tmat...@google.com >> > -- > 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/-/7AtnmRTCRToJ. > > 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 | Developer Advocate | tmat...@google.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.
