Node-RED inaccessible 10-15 minutes after start up

General
1

Hi All,

I just started loosing access to the web interface sometime after Node-RED starts. All the flows still appear to be running and the logs show no indications of anything bad happening. When working I can access the service at https://127.0.0.1/1880
Anyone know what could be going on?
After a restart it come back up for a short time.

2

Are you running on windows?
Are you running node-red behind a proxy?
Can you show us the log from start up to the first message after this delay?

3

Sorry I have should have included some more detailed info.
Yes running on windows.
No proxy.
I am away from the server so will have to get the log tomorrow.
Node-RED v4
Node JS 20.something

4

Also check the process monitor, or whatever it is called in Windows, to see if node red is hogging the processor when it stops responding.

5

Here is the start of the log.

Welcome to Node-RED
===================

30 Aug 10:29:37 - [info] Node-RED version: v4.0.2
30 Aug 10:29:37 - [info] Node.js  version: v22.7.0
30 Aug 10:29:37 - [info] Windows_NT 10.0.17763 x64 LE

Even with Trace on there was nothing in the log about what was ran or was running when the interface went down.
Looking at the task manager when the interface was down NodeJS and Windows Defender had higher CPU usage than normal. Unfortunately I don't have high enough privileges to turn of Defender.
With every flow and node disable it became stable again. Slowly starting to turn things back on...

6

Node v22 is not yet supported, so it may be that. Drop back to 20 and try again.

https://nodered.org/docs/faq/node-versions

7

Was on 20.17 and tried 22.7 to see if it made any difference.

8

OK, so not that. I still suggest going back because some nodes do have problems with 22.

When you said the CPU used by nodejs was higher than normal, what does it normally show and what did it change to?

Are you using dashboard charts? If so then disable those and see if it still happens. If that fixes it then over what time are the charts showing, how many lines on each and what sample rate? Add up the total number of points in the charts when they have filled up and see if it comes to tens of thousands.

9

Node JS goes from idle 0% to 5-10% CPU
Defender goes to 20-30% CPU

Dashboard is being used and so far all charts and graphs are re enabled and working.

10

Did you work out what was causing the problem?

11

No not really, the system was stable over the weekend but I did not redeploy all the flows yet. We have NR running on a production and a development server so I may try copying over our current prod flows and seeing what happens.