WTH, are Dashboards are getting resource intensive? Who knows, but sometimes it feels like it

I have about 50 views divided up in 4 dashboards. At times the dashboards appear to be super slow and difficult to impossible to use. While the causes surely differ (one big offender is RTSPtoWeb - webRTC sucking up over 50% of my i7’s CPU, at least until recently), I am starting to wonder how to preoperly design dashboards to lighten up the load and ensure responsiveness.

Recently I saw a note on Card Mod saying that if you prepend card_mod: before Style: it would use less memory. That got me thinking… what else could there be that may be affecting my system?

It would be nice to have some metric to understand the impact of each view in a dashboard, or a single dashboard, or all dashboards, whatever is feasible, to understand whether the system load / responsiveness is being negatively impacted by how it is built, and/or changes made.

That would be a cool dev tools addition

1 Like

@aruffell Just started to use WebRTC as well with 6 cams. 20% more CPU usage on my HASS machine. You mention “until recently” ; Did you stop using it or dud you find another way to reduce CPU usage?

@Tyfoon - I believe, but I am not sure, that there was an improvement after one of the latest updates. I don’t know whether it was the update itself or something else but my CPU usage appears to have dropped.

Right now it is around 22% with 10 IP cameras configured in HA. Most are 8MP and the NVR is set to always send the maximum resolution and bps stream as it did before the drop.


Yesterday I was working on a 12 channel ESPHome energy meter and set it to 3s update interval to more easily calibrate it. I forgot it set at 3s and later in the day I noticed that the dashboards were extremely sluggish. I found a log entry about “Client unable to keep up with pending messages” and “The system’s load is too high or an integration is misbehaving” which led me straight back to the energy meter. As soon as I toned it down to 10s, all was well.

So, I have 2 confirmed cases where integrations were sucking the life out of HA making the dashboards super sluggish… however I’d like to know whether what I do on the dashboards is contributing to the load, and if it is, how badly…

1 Like