Theoretical maximum value of <max-concurrent-requests> for Google AppEngine Backend

Question

does anyone know whether there is a maximum value for the <max-concurrent-requests>-setting for AppEngine Backend Instances? I.e., will a backend be able to accept and service requests until it entirely runs out of memory or CPU, if I set <max-concurrent-requests> to a giga-billion? Or are there other quotas or external limits imposed? (I'm thinking TCP port numbers, thread handles, ... ?)

Answer 1

From https://cloud.google.com/appengine/docs/python/config/appconfig:

max_concurrent_requests:

Optional. The number of concurrent requests an automatic scaling instance can accept before the scheduler spawns a new instance (Default: 8, Maximum: 80). You may experience increased API latency if this setting is too high. Note that the scheduler may spawn a new instance before the actual maximum number of requests is reached.

Answer 2

Quoting from documentation of GAE

"Specifies the maximum number of requests that each instance can handle simultaneously. If unset, App Engine determines the concurrent request limit dynamically."

So they have not specified the limit, but also have indicated that GAE will determine the limit dynamically, which might be based on number of instances you have and what kind of instances you have configured. The ability of those instances to handle your requests will probably determine the max.