Hi all, I too have seen strange behavior following the datastore issue (#13551) - and this is going on every since.
Instances are a bit crazy going up and down, I've seen latency spiking (i.e. 20 seconds for a task that takes < 1sec and is datastore/memcached bound), memcache seems kind of "exhausted", tasks staying in queues for 2 minutes, etc. All this behavior is definitely *very* different than before #13551 (which may be irrelevant but I'am mentioning it only as a time milestone). Can you please let us know if you're experiencing issues with EU region? On Wednesday, February 22, 2017 at 9:37:16 PM UTC+2, Linus Larsen wrote: > > After some digging I found out that we've made some modifications on the > mobile app, which makes it do several ajax requests rapidly. > The result became somewhat strange. > > 1. Request1 starts a new instance > 2. Request2 also starts a new instance since it takes aprox 20 secs to > start a new instance, and the instance started in request1 isn't ready yet > 3. Request3 also starts a new instance since the instance started in > request1 isn't ready yet > > Suddenly we have 3 instances running, and the response time wasn't > improved since the startup time of each instance it so slow. > I guess we can improve the behavior fiddling a bit with max/min-pending > and min-idle instances settings. > > But. As it seems, started instances are shut down very quickly, more > quickly than they used too. > > FYI, it all started last week (and it not just me who noticed these > changes) > > https://code.google.com/p/googleappengine/issues/detail?id=13551 > > And it keep on going with different kinds of issues, today pub sub calls > timed out, memcache was totally down, datastore failures etc. > Very shaky, and it makes me wondering if Google is doing som changes in > the EU region. > > > > > > > > > > > > > Den fredag 17 februari 2017 kl. 16:49:29 UTC+1 skrev George (Cloud > Platform Support): >> >> Hello Linus, >> >> You seem to have 3 projects running at this time. Which one of the 3 is >> affected by the 100 instances for no reason at all behavior change? For >> confidentiality, you can send us the project name by private email. >> >> None of the 3 projects start more than a maximum of 17 instances at a >> time. When did this change in behavior occur, exactly? >> >> Looking at your versions in the console, one notices that "automatic" is >> chosen as value for the "Min Idle Instances" and "Max Idle Instances" >> parameters. >> > -- 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/dff4d6b1-3162-467b-95c9-c40ca52d564e%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
