Weird scheduler behavior (again)

[Francois Masurel]

Could someone from Google explain why a new dynamic instance was started though an existing one was still alive a served only 1 request for the last 15 minutes ?

Really looks like a waste of resources (and money).

Thanx for your answer.

François

[Takashi Matsuo]

Hi François,

The basic rule to determine a need for a new instance is described here:

http://stackoverflow.com/questions/11525717/when-does-the-app-engine-scheduler-use-a-new-thread-vs-a-new-instance

Other than that, there can be some rare-case scenarios where the scheduler need to spawn a new instance due to some transient issues within the underlying infrastructure, which are not surfaced to developers, and handled automatically by the scheduler. I think it's a nature of a distributed and shared system like App Engine.

I assume you already have starred the following issues, but please star them if you haven't.

https://code.google.com/p/googleappengine/issues/detail?id=7865

https://code.google.com/p/googleappengine/issues/detail?id=5858

--
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/-/wZs7GyY8z1sJ.
To post to this group, send email to google-a...@googlegroups.com.
To unsubscribe from this group, send email to google-appengi...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/google-appengine?hl=en.

--
Takashi Matsuo | Developers Advocate | tma...@google.com

[Francois Masurel]

Hi Takashi,

Thanx for trying to help.

Is there a particular reason why all this information from SO isn't in the official GAE documentation?

Could you explain a bit more how the instance class CPU limit can be reached and what we can do to avoid it?

Thanx again for your help.

François

PS:

All issues have been starred long ago.