Any CPU and RAM generic metrics for deploying django with gunicorn on a aws arm instance (t4g) ? | Django | Page 1

shell whale Aug 19, 2023, 1:48 PM

#

I am trying to figure which instance to go with in AWS (cpu and ram config).
the load will be around 2-5k users and is heavy on database calls (aws RDS postgres), there is minimal computation. Since the server will have 2 cpu, number of works will be 5 (as per docs formula 2*cpu + 1), but I am unsure oh how much ram to go with, either 1gb or 2gb.

rare lintel Aug 19, 2023, 2:00 PM

#

I would start with 1gb but measure and monitor usage. Then upgrade when needed.

I don't think it can be predicted

shell whale Aug 19, 2023, 2:34 PM

#

right I was considering that but in general, was looking for some public metrics
will start with 1gb instance and increase if needed 👍

rare lintel Aug 19, 2023, 2:37 PM

#

There's k6.io to do load and stress testing

shell whale Aug 19, 2023, 2:40 PM

#

oh cool, will check it out, thanks

rare lintel Aug 19, 2023, 2:43 PM

#

Once you find a ec2 instance that works, I'm betting the bottleneck will be the database

shell whale Aug 19, 2023, 2:49 PM

#

rds is already running on t2 micro and it usually gets to 15% cpu utlization on the old nodejs project, but django will defenetly use up quite a lot more, need to monitor and figure out if that needs an upgrade as well

#Any CPU and RAM generic metrics for deploying django with gunicorn on a aws arm instance (t4g) ?