Answer the question
In order to leave comments, you need to log in
Setting the number of running instances
In the application settings, you can select max idle instance. If the value is 1, then is it possible to expect that all requests will go to one instance and, accordingly, use its RAM as a cache protected from collisions.
In my case, I am not afraid of the periodic data loss associated with the collapse of the instance, but I must know for sure that the data will not be distributed among several instances. Can I count on it with these settings?
Answer the question
In order to leave comments, you need to log in
The help says that in some cases the actual number of instances involved may exceed the maximum number of pending instances from the settings.
So the answer is no.
No, from experience I can say that either 100% can not be said for GAE. Firstly, it is impossible to control the lifetime of an instance, and if no requests are received, then GAE kills the instance. Secondly, even if you pay to keep the instance hot, it often happens that GAE restarts the instance (I suspect that the memory does not leak). It's better to write everything in threadsafe, and in your case, use memcache.
PS but in instances that backend live for a very long time, and there are almost no limits that are inherent in ordinary front-end types.
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question