As a deployment expert, I'm often asked questions like "Which server do I need for xxx users?" or "How many requests per minute can my server handle?". 

This talk will explain how we answer those questions, the methodology we're using to answer them, and how to interpret results.    

We'll also answer the question "is it better to scale horizontally or vertically?".  We'll show our tools, as well as our results; plenty of beautiful graphs :-D.  

We've conducted a load test which leads to 3k requests per second handled by a single Odoo server. We'll show our results, how we reached this point, the pain points we faced, and more.  

We'll also show the differences between an on-premise Odoo and an odoo.sh instance, with the same code.  Finally, we'll give you advice on how you can conduct these types of tests, and why it's important, or not, to do it before a go-live or before any major release.