Is the API restarted on every request?

[s4m-mo]

I make a request to my API, and for every single request, the micro seems to be starting up with the opening Uvicorn logs.

My question is, every time I make a REST API request to my Deta Space Micro, does it run the micro from scratch every single time?

[SlumberDemon]

Yes, A space app only stays alive for 20s which is its timeout and after that timeout it gets destroyed, this is on a request based system so this only happens when a request is made to that endpoint. This is because deta is a serverless platform.

More on serverless platforms can be found here

https://www.cloudflare.com/learning/serverless/what-is-serverless/

[s4m-mo]

Thank you for the clarification. I have a couple of questions.

• Will any future paid Deta plans allow for a greater timeout?
• Also, do you have any tips on improving python FastAPI startup times?
  Could I run a python py_compile task on space push and then execute the .pyc file rather than the interpreter having to re-bytecode-compile the source every time?

[SlumberDemon]

Will any future paid Deta plans allow for a greater timeout?

There are no plans for anything like that yet

Also, do you have any tips on improving python FastAPI startup times?

Not really

Could I run a python py_compile task on space push and then execute the .pyc file rather than the interpreter having to re-bytecode-compile the source every time?

You can't skip any steps during the push