GAE first generation memory size 128MB or 256MB?

[KWagdy]

Hello,

According to GAE Docs older version of Python 2.7 GAE applications has 128MB for F1 instances, yet we see on the instances memory graphs that the average memory consumptions lies somewhere in the 350MB range, has all python versions been upgraded to 256MB F1 instances, i.e. it's already on second generation runtimes?

I have seen this discussion on the old Google Groups, but it doesn't seem to have a definitive answer, anyone knows definitively?

[christianpaula]

Hi @KWagdy,

Welcome to Google Cloud Community!

The first-generation Google App Engine (GAE) instances have a default memory size of 128MB. However, the actual memory usage of an application can vary based on factors such as the size of the application, the number of requests it receives, and the amount of data it needs to store.

It is also possible to configure the memory size of GAE instances manually, which can result in instances with a larger memory size. However, this will increase the cost of running the application.

Regarding the upgrade of Python versions on GAE, it is recommended to check the release notes and documentation of the specific version you are using to confirm any changes to instance memory sizes. It's also possible that any increase in instance memory size is due to the second-generation runtimes, which offer more flexibility and customization options for GAE applications.

Thank you

[KWagdy]

Hello @christianpaula,

Thank you for your reply, but unfortunately it is still yet not clear to us. The reasons we are asking are:

1. We see an average of 350 MB of memory used, sometimes more, yet the number of instances seems consistently at 2 active instances (we have a few number of users). If the RAM per instance is 128 MB, then they total to only 256 MB, which doesn't suffice the average of 350 MB as noted earlier. 
2. We received some notes from the users that sometimes they get the browser error "ERR_CONNECTION_TIMED_OUT, <website> took too long to respond" which we are trying to debug. We suspect that this might be due to the F1 instances not being able to handle the requests with their current specs. So understanding the specs might help us in our investigation.

But since you are a staff member, would you be able to get a definitive answer in this regard internally? Would be highly appreciated, 

Thank you,