Node-RED as a platform for complex web applications?

[Meloman]

A few years ago there was a discussion here about using Node-RED as a CMS.

"Is a CMS built using Node-red even a good idea?"

Well, there are lots of CMS options that are easy to set up and efficient to run. A 1.5GB AWS instance is pretty big. I used to run my own blog in WordPress in 1/4 GB I seem to remember and WordPress itself is considered quite heavy as it needs MYSQL as well. These days, my blog is run from a static site using Hugo. It is hosted on a free-tier from Netlify with Cloudflare (again free tier) providing a front-end proxy that gives some additional performance, security and analytics. I'm only getti…

I think the more interesting question is slightly different: is Node-RED a good platform for building process-driven web applications?

I've been using it this way for some time now, including our company's public website and an intervention-management web application, both running in a DMZ.

I use my own flows for user management and application structure, with Liquid templates for rendering, while Node-RED handles business logic, databases, APIs and workflows.

As a real-world example, the intervention-management application currently gets:

GTmetrix Grade A — Performance 92% — Structure 91%
LCP 1.6s — TBT 0ms — CLS 0.01

Personally, I wouldn't use Node-RED as a traditional CMS. But for complex web applications where the complexity lies in business processes and integrations, I find its flow-based model particularly well suited.

I'm curious how many others are using Node-RED this way, and how far you've taken it.

[gregorius]

It was actually one of the first things I created with Node-RED: it was called RedBack - Node-Red as Backend and lead to one of my favourite comments here at the forum:

@tree-frog

What your looking to do is reinvent the wheel with the same wheel. What I would suggest is working with Node-RED as it is (Which is pretty dang great), and get familiar with what it is and does, then you can do amazing things.

So it has the potential to get misundertstood when using a web application to create another web application, even if you think it's amazing to create a visual web application with Node-RED.

EDIT:

I continued on and my blog on Node-RED is running NR with the content actually being the flows, i.e., articles are flows, there is no database backend to store content.

[Aaqu]

For some time, I had been looking for a system that would let me handle devices communicating over different protocols such as Modbus, store the data in a database and visualize it all in one place.

Node-RED has worked very well for me in this role. Dashboard 2 did not really fit my needs, so I started building my own frontend module based on React.

At this point I already have more than a dozen different tabs and applications running in production and their use cases are quite different. Some are only used to visualize data from MSSQL, while others are based almost entirely on direct communication with devices. They read parameters, change setpoints, set temperatures and control parts of the process.

What I also like is how flexible this approach is. Today I need one feature, next week someone asks for another one and then something else needs to be changed. I just open the React editor inside the flow, make the changes and click Deploy.

With a typical standalone React application and a separate backend there is usually more involved. You need to build the application, create a new Docker image, replace the image, restart the container and so on. Of course this can all be automated, but in my case Node-RED simply gives me a much faster development and deployment cycle.

That is why it works so well for this kind of application. I do not really see Node-RED as a traditional CMS. For me it is more of a platform for building process driven applications where a large part of the logic is related to devices, databases, communication and control, and where requirements can change very quickly.