Node-Red via VRM "stopped" working upgrading to OS v3.50

[Anders (Anders Røil)]

I am running a Cerbo GX with Node-Red and I am daily using Node-Red dashbord to remote control my system .

After upgrading to v3.50 I experienced that Node-Red accessed via VRM became very unstable and in practise useless. It lost connection after seconds.

I also lost some of the data point I to pull from the VRM API, i.e forecast values. I assume that issue is related to the issue i mentioned in this Topic, but I haven’t checked.
[VRM API - ID number change because of VRM update?]

Anyone else with a similar experience?

Reinstalling v3.41 solved the issue, it also brought back the VRM API data

This small calculation is made based on these VRM APID Forecast data.

Yes - it’s in Norwegian (partly - sorry) but your probably get it. I have 1.12Kwh available to “spend” the next 24h

[guystewart (Guy Stewart (Victron Community Manager))]

Thanks for reporting @Anders - I’ve asked the team to have a look

We aren’t able to reproduce this yet, if anyone else is experiencing the same issue or has resolved something similar in the past please reply here.

[victronvalentin]

Good afternoon,
I contacted you yesterday on the email address listed in your profile to try and diagnose this issue further.

Could you check your emails and get back to me?

[Anders (Anders Røil)]

You have a reply from me now

[victronvalentin]

Thank you, I have responded and need your answer.

[victronvalentin]

Good evening,
After looking into your system with your approval by email, I do believe that the NodeRED issues you’re experiencing are due to the fact that you reboot the NodeRed service when the CPU is higher than 60%, as you told me in your email.

I would suggest to disable this or optimize your NodeRed code.

I do not see anything else strange in your system that would explain those issues.

[Anders (Anders Røil)]

Just a final comment from my side on this topic to keep you all informed.

My node red code is slowly consuming power due to bad coding. This I have sorted by using the “killall node-red” if CPU increases over 60%. However since the new UI consumes substantially more CPU power the 60% limit was hit immediately. The drop in below chart is when I turn off the new UI.

I have increased the threshold to 90% and this solved the problem. I obviously still have to to sort out my coding issue.