Im conceptualizing a controller, to manage a industrial battery, in order to dispatch it in high peak demand depending of the solar generation. This is an aplication that should be running 24/7/365, All the communication with the devices is in modbus protocol. I need to storage all the data in a data base, and them see it in a platform like grafana. My question is: Is nodered a good option con control the devices 24/7/365, or should I thing in another solution to integrate the control? In case you thing nodered is not the best option, which do you thing is the rigth hardware/solution?
As a platform, nodered is extremely stable. So, it is a good solution.
HOWEVER, there is no such thing as an application running 24/7/365.
If you really want that, then, running (at least) two instances of the application in an active/active or active/passive mode is a must.
With two power supplies...
Redundancy on every possible component of the entire solution is a must.
At a bare minimum, you will have to update the system or elements of the production environment. Hence downtime.
So, NodeRed, yes. Don't forget about everything else.
Yes Node Red can run 24/7x365 - but you have to cater for upgrade and patching paths in your deployment.
In my case i run my Node Run on a cluster of Virtual machines, when it is time to perform an upgrade on the underlying OS and/or Node Red, i clone the Virtual machine (ESXi so no downtime)
I bring up the newly cloned image on a different isolated network (which has no access to my production network and hence no collisions or tries to control devices/peripherals on my network) i perform the upgrade as required and ensure the system restarts in a stable fashion. Once this is all done i then schedule a time for the changeover from the live running system to the newly patched system.
Once ready for the changeover
copy the running Node Red config file from the live system onto external media, download into the updated VM, Restart Node Red,. Run a flow/Bash Script to copy all relevant files (logs etc) onto a centrally accesible system.
Then - This involves about 10 seconds of downtime, Change the network assignment for the live system to a different network segment and change the clone VM network assignment to the live network segment
Run a node Red flow that i have to copy all the additional files as required back onto the now production box
I have the two ESXi hosts on two different UPSs as well as on a pair of network switches that are "stacked" together.
It is only my peripheral devices that are not redundant (such as my Modbus/RTU to Modbus/TCPIP gateways (but i keep a hot spare) and download the configs from these once a week to a central location.
I use my node red systems to monitor and manage my solar and battery system and all the discretionary loads in the house
