เมนู
Load Testing: How to Stress Your Odoo with Locust
Developer
ที่ตั้ง: Aula Magna / Lauzelle - 4/10/17 14:30 - 4/10/17 14:50 (Europe/Brussels) (20 นาที)
Load Testing: How to Stress Your Odoo with Locust
Michael Vannerom, Developer, Odoo
Michael Vannerom, Developer, Odoo

Michael is working for Odoo for 6 years as a technical consultant. During this period he has the time to see a lot of different topic for example: performance issue, giving training, reporting, make some big project for Odoo customers,....


Load testing: How to stress your Odoo with Locust

Deploying an Odoo instance on production can be an adventure without its challenges. If you want to get the maximum resources from your server, you have to adapt different options. It is not always easy to know the number of workers and the different limits depending on them. But after you setup your server how do you know if your theoretic value seems to be correct and when can you see the first error appear. To be sure that your system stays usable it is important to find a way to stress your Odoo instance with a different number of users.

To answer all these question we will show you what we develop internally with Locust. Locust is an open source load testing tool that allows you to send XMLRPC requests with a specific number of users.

Just imagine the different scenarios you want to test, then with this library you can write a test in Python with the lib XMLRPC. After that you have the locust interface just to say I want to test with 1,20,50, 100,... users. It will prevent wrong configurations and check how many users the system is able to handle while still answering requests in a correct time. You also have interesting stats per request like the number of failing requests median time, average time, min and max time for a request and how many request per sec Odoo is able to answer.

With all the aforementioned statistics you will have an accurate view of the limits of your system.

During this talk we will see how to set up your locust on your system, case by case. To understand how to use it in a practical case we will also show you a real case and how we use it internally. After this talk, attendees should feel more efficient and you will know the limit of your system.