The comment about Node-RED editors is specific to FlowForge Devices as run with the @flowforge/flowforge-device-agent, those instances of Node-RED will show a login page for the editor.
This is NOT instances started from the FlowForge web interface.
That image without a lot more context doesn’t help. But the localfs build really is not intended to be run behind a nginx proxy, because all the URLs for instances will be port numbers on localhost, not a proper hostname.
I think we need a full description of how you’ve set things up to know if it’s going to work.
The docker/kubernetes based version is intended to be hosted behind a proxy and makes use of proper hostname based routing.
nginx proxy with flowforge behind it => This is not good as I understand it, it should be without proxy or we should use docker.
Edge devices as Device Agents. They have been configured, but they already ran Node-Red before. As I understand it, this will conflict with the device agents, and I should first uninstall Node-Red from the devices. When I visit the Node-Red page of one of my edge devices, that also has device agent installed, I get the ‘Cannot GET /’ screenshot. I found it a strange error, so I was wondering if this had something to do with the MQTT missing, or if it’s just because of the Node-Red missing…
OK, the image now makes a little more sense. it will depend on what version of the device-agent you are running,
Older version do NOT run the editor, you are supposed to edit the flows in the cloud instance and push them down to the devices.
In the latest version of the device-agent, the editor is running but you will not be able to access it directly, you will be shown a login screen, but there is no way to directly log in.
Accessing the editor is a enteprise feature that requires both the MQTT broker to be installed and an Enterpise License. Once enabled you can access the devices editor vai the FlowForge web application which opens a proxied tunnel to the device and handles all the authentication. This is intended to allow for debugging when the device has access to local only resources, it is not intended to be used to directly write the flows.
Personally, not being able to debug in the edge editor, makes the open source version quite useless for me.
Having 50 Node-Red instances connecting to different machinery, I want to know if my code might not be working for one of them, but not being able to debug fast on the edge device is a big turnoff.
Will this feature be opened up for open-source in the future? Else I’ll stick with a self built solution to handle this.
There’s no plans currently to port this feature. FlowForge is an open core product, and we identified the is most useful for large scale adoption like yours. Large deployments are where the licensed edition is most valuable, we’re happy to discuss with you how to deploy FlowForge Premium for your use case.
Note that the logs of remote Node-RED instances are available in the OSS tier of FlowForge.
I recently posted a question about real time monitoring and debugging on the edge gateway from the cloud. This I understand isn’t possible. I also maintain local connections with devices in a local network, this also doesn’t look possible to look up other devices on the local network from the cloud environment. So, for me this isn’t an option.