We’re running Squidex on Azure and are running into performance/timeout issues. It’s deployed on a single Linux App Service Plan, S1 (1.75GB), using Docker.
E.g. just tested adding an asset with two 1MB assets, and memory usage went up to 100%, and got an error. Average response time went up from ~30 ms to about 4 minutes.
1.75GB is not a lot, but why doesn’t this seem to be enough for 1 simple request?
What are the recommended specifications as I can’t find anything in the documentation about hardware requirements?
There are no specs defined, I just don’t know. But 1.75GB including MongoDB is probably too low. Especially when both are running and fighting for the same resources.
Thanks @Sebastian. We’ve got Mongo on a different server. But still 1.75GB for just the Docker container doesn’t seem enough.
I’ve now increased it to a P2v2 (7GB) which seemed to have solved the issue.
What would you recommend as minimum specs? Should we scale out as well? Or is 1 node enough?
1 node is enough. I will benchmark the memory usage on file uploads. I do not buffer anything, so it should be okay, but lets see if this is really the case.