I just installed a self host instance on the Digital Ocean Apps platform, following the instructions here.
Everything was pretty straightforward and easy to setup, but running into an issue with the memory staying incredibly high. I first installed it on a container with 1vCPU and 2GB of ram, according to the instructions. The memory remained at about 95% despite barely using the app, there is clearly memory leak or some other issue, but every setting was based on the instructions so at a bit of a loss. I tried increasing the server to 4 GB RAM | 2 vCPUs but still in the high 70%'s even when the app is not in use.
Has anyone had a similar issue and know what needs to be changed to fix this? Or any ideas on how to narrow down what is causing the issue?
Hey @bennyv53, this could be related to the way Python works. We’ve observed it in our hosted offering as well on baserow.io as well. When initially using the tool, certain parts of the codebase are not used, and therefore not loaded into memory. While using the tool, these parts could be used for the first time, and would use a bit more memory. We recommend a minimum 3GB because that’s what it can grow to with a basic setup, but it should stop at some point.
You can control the memory usage by setting the following environment variables:
By default, these values are set based on the number of CPU cores. If you lowerr then, it does mean that less HTTP requests can be handled concurrently, and less async handled concurrently. If only you are using the Baserow instance, and you’re making automated API requests, then it should be fine to set it at 1.