Thanks Steve, yes happy to chat with anyone here regarding how FlowFuse can manage all of this for you.
Here is a link to my calendar: https://meetings-eu1.hubspot.com/flowfuse/zack
So you suggest not to use a physical machine, but to get a cloud workstation? We don't need a real "server" for this, as we have no need for the usual server services (this was actually adviced by a consultant)
The important part, is the end, is to have a dedicated device for node uses. Understanding specifics is what I aim to with this thread 
Well, I already mentioned Linux. And even without much Linux knowledge, if you want a "workstation" for running node.js microservices, it is hard to beat. Maybe have a go with a simple Debian "headless" VM to see what I mean. Bearing in mind that I have over 40 years experience of IBM mainframes, OS/2, LAN Manager and Windows and I still would always reach for Linux for a simple server which is what would be ideal here. In any case, with the advent of AI such as ChatGPT 4o, explanations of command lines is mostly a simple query away.
The points for using a cloud VM:
- You can adjust their sizing easily and instantly to cope with peaks and troughs of need.
- You have no "tin" to have lying around wearing out and using power and getting clogged with dust. Not to mention, no possibility of the cleaners unplugging it to use the vacuum (and yes, I've plenty of examples of that happening even in comms rooms).
- In some cases, you might not even need to maintain the OS.
- You can easily automate turning it on/off and this will save you money if there are parts of the day/week/month where you don't need it at all.
- It's "design" can be easily templated so creating a new instance is a 30sec task. Meaning that you can easily automate dynamic scaling.
- Assuming that the VM(s) are in the same AWS data centre, you won't get any exfiltration costs.
I think it wouldn't be a problem to take a linux machine, if you think that is the way to go. It will need some adjustments for the migration I think? for instance the exec nodes should have the linux cmd commands, right? But I think it's still manageable.
Generally speaking, what specifics would you recommend?
talking about CPU, RAM, number of disks, stuff like that.
Yes.
Sorry, hard to know without knowing what scripts will run. I'd start with something small. After all, we can run a basic Node-RED setup on an embedded microprocessor so anything that will run Debian headless should easily be enough to start. Then you will need to do some testing anyway for your migration of the scripts. So that should give you a good feel for things.
Just make sure to use AWS templates (sorry I know more about Azure than AWS so not sure what they actually call them) to create the VM so that everything is nicely repeatable.
