Open BeritJanssen opened 2 years ago
Do we have any information about the CPU / RAM usage of the VM, about the percentage of successful / failed requests, and about the average response time?
We can of course inspect the servers, but I don't know if there are any automated warnings in place if CPU/RAM usage peaks. If not, could we set something like that up via Sentry, too?
Judging by this article it is certainly possible to do some performance monitoring. I think this would be primarily for Django API endpoints then. For static files it might be a bit trickier since Sentry (in the browser) might only initialize after other files have been loaded already.
I do think that Django is the bottleneck here - most of the heavy duty is done in the backend.
If we have multiple popular experiments running, will servers still handle traffic?