In my system, when there is no browser access to the gui v1, its process is consuming about 2% CPU.
If there is a browser accessing the main screen of gui v1, the gui process is consuming about 4% CPU.
Your 18% seems a bit off…
More of that it’s memory consumption stays for hours the same, web access or not.
I have a system working for 180 days now and all is stationary and there were web accesses alright.
But my Venus OS is version 3.32, so it may be of no use for your situation.
I followed the steps he mentioned above and I was able to reproduce the issue. I am on 3.66 with MQTT and a custom MQTT Temp sensor. When I start GUIv1 and make a change and leave it open the free memory drops until I close the page. When I start up the page again the memory start to drop again. I had stable memory before this. The red arrows indicate where I opened GUIv1. As soon as I close it the memory stabilizes, but does not return to free memory. I use the same SSH integration method as Ingo to graph the free memory.
Hey both, great to have this cleared up! thanks!
I’ll discuss this over here - to decide if we’ll look into this. A reason why not is that we’re primarily focussing on gui-v2 - hope you understand.
And this bug is then not something in the application that we develop; its in the libraries and existing tooling we use for this Remote Console feature.