Hmm, it's hard to tell what might cause the issue from this level of abstraction, but if you ever felt like it, you could produce a small reproduction and a technical report and pass it off to the Public Issue Tracker <https://code.google.com/p/googleappengine/issues/list>. It could be a complex case of WorkAsIntended, or it could be exposing a little-realized issue, the only way is to investigate deeper!
On Tuesday, January 19, 2016 at 4:58:00 PM UTC-5, Kaan Soral wrote: > > Scenario A > 1) Fetch 40 entities > 2) Fetch an external url, takes 1-15 seconds > 3) Put 40 entities async > > Scenario B > 1) Fetch 40 entities > 2) Internal computation, no delay > 3) Put 40 entities async > > So the issue is seldom at A, yet it floods the logs at B > At both scenarios, multitude of these operations are ran, 40M entities for > B, <<1M for A > (A is an hourly operation, while B is more of a one-time thing) > > Since there is no observable data/operation loss, I didn't debug the issue > > On Tuesday, January 19, 2016 at 11:41:46 PM UTC+2, Nick (Cloud Platform > Support) wrote: >> >> Hey Kaan, >> >> Interested to try to analyze the source of the issue - what do you mean >> by "if the routine is slow" - does this refer to the cron schedule being a >> larger interval, or the time for requests to execute? Are you able to run >> memory diagnostics in-request to log the rate at which memory gets used-up? >> >> On Friday, January 15, 2016 at 6:08:49 PM UTC-5, Kaan Soral wrote: >>> >>> Interesting, thanks for the link >>> >>> The app I mentioned used db, my new app uses ndb tho I don't experience >>> memory issues frequently with it >>> >>> The app that experiences memory leaks has crons that pulls and processes >>> 40 entities at a time, entities are <1MB, so when instances are flooded >>> with these crons, memory issues occur frequently >>> Interestingly, if the routine is slow, they occur less, but if the >>> routine is fast, it occurs every 30-40 requests >>> >>> On Friday, January 15, 2016 at 9:42:23 PM UTC+2, Christian F. Howes >>> wrote: >>>> >>>> FYI, if you use NDB there is a memory leak: >>>> https://code.google.com/p/googleappengine/issues/detail?id=9610 the >>>> google folks are having trouble pinning it down, but it's there. i just >>>> let my instances restart as they run out of memory as i have not been able >>>> to find my other memory leaks (i'm sure i have more then just the NDB >>>> leak). >>>> >>>> On Friday, January 15, 2016 at 9:50:24 AM UTC-8, Kaan Soral wrote: >>>>> >>>>> I have an app that kills instances after just 30 requests sometimes, >>>>> it functions perfectly tho, the instances die pretty gracefully, my >>>>> assumption is that they don't take anything with them, rather handle all >>>>> their requests and die after that >>>>> >>>>> I don't upgrade to F2 as it increases the costs significantly, it's >>>>> harder to utilise F2 instances efficiently >>>>> >>>>> My .02 >>>>> >>>> -- 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 https://groups.google.com/group/google-appengine. To view this discussion on the web visit https://groups.google.com/d/msgid/google-appengine/282f39b4-930e-4e56-8368-86b22f9c4cb3%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
