gRPC

Setup

Install the dependencies:

$ python -m pip install --upgrade pip
$ python -m pip install grpcio
$ python -m pip install grpcio-tools

Performance

Everything below (only gRPC) don't make a lot of sence because sometimes CPU goes to 100%, sometimes only 60% and after few reruns 100%

Hello <Name>

time displays 1000 responses of Hello <Name>

	1 client	10 clients	50 clients
1 worker	0.4-0.53 sec	0.32-0.43 sec	0.34-0.5 sec
2 workers	0.41-0.45 sec	0.34-0.40 sec	0.34-0.37 sec
4 workers	0.4-0.6 sec	0.34-0.36 sec	0.34-0.41 sec

Add user

time displays 100 added users

	1-50 clients
1 worker	0.72-0.83 sec
2 workers	error

Auth user

time displays 1000 auth requests

	1 client	10 clients	50 clients
1 worker	2.67-3.5 sec	2.67-2.77 sec	2.95-3.3 sec
2 workers	3-3.25 sec	3.1-3.5 sec	3.5-5.1 sec
4 workers	2.42-3.3 sec	2.74-3.6 sec	2.5-3.2 sec

Conclusion

GIL exists, time of all operations with database pretty random, sometimes it's 1.8 sec, sometimes (same code, another run) time is 6 sec.

FastAPI

Setup

Install the dependencies:

$ pip install fastapi
$ pip install "uvicorn[standard]"

Usage:

$ uvicorn server:app --reload --log-level 'critical'

Performance

	async client
1000 Hello	0.50-0.56 sec
100 Add user	0.17-0.34 sec
1000 Auth user	2.18-2.60 sec

Flask

Make no sense

Performance

	async client
1000 Hello	1.35-1.60 sec

Django

Make sense because ready made admin + a lot of user libraries for literally everything. Read performance sucks idk why 😔🥺

Performance

	async client
1000 Hello	0.95-1.16 sec
100 Add user	0.50-0.57 sec
1000 Auth user	6.16-10 sec