Please suggest server specs for a large scale Phoenix application

Hello,

I need advise on the best approach to deploy my Phoenix project. It is recruitment portal that will receive a huge number of applications from the first day to run for a period of six. Past experience has shown that we will get up a million applications from the go.

Our plan is to use Linode VPS with 6 CPU Cores, 12 GB RAM, and 8 TB Transfer. What would you advise? multiple servers? how many and what configuration? Load Balancers. I just need general advice.

Thank you

1 Like

I would always suggest a dedicated server if your budget allows :slight_smile:

Check out these threads too:

And there’s quite a few dedicated server companies mentioned in this one:

The deployment tag may be of interest too :smiley:

Start with this one, simulate the load that you expect, then maybe pick a bigger/smaller box and repeat the simulation. The good thing about virtualized environments is that you can add more boxes like this later if needed very quickly.

3 Likes

Thank you @AstonJ I will read those threads

1 Like

Thank you. How do I simulate the load? What tool can I use to achieve that?

1 Like

You can use tsung.

3 Likes

Both tsung and wrk are good. I tend to prefer wrk for higher loads (it is far more efficient), and it uses lua for scripting.

I would recommend Gatling for more real benchmarks, simulating real user actions instead of raw performance

1 Like