Hey @masterofweb,
It is going to be tricky, Because n8n does everything in memory it is going to be a case of increasing the memory available to the container / node process so it could need something in Docker unless Caprover has an option.
The downside to this though is you would probably need to keep increasing it to find the value that works for you and eventually you may hit the limit for memory you have available. I know you don’t want to do it but it might be worth starting with splitting your workflow up.
@MutedJam do you have any thoughts?