Continuously casting dashboards, does it work for you?

I’m just looking for some insight here. I am trying to use the continuously casting dashboards hacs integration, but it doesn’t work very well for me.

Does it work for other people? Like their dashboard isn’t laggy and the dashboard is cast to the Google display for more than a few minutes?

I have the ha-catt-fix installed and loaded into Lovelace resources. The ccd logs show that the “Dummy” stream (from catt-fix) is indeed being detected. I tried increasing cast_delay and restarting Hass and the display with the same results.

My Hass and Google display are on the same physical network while testing, so no multicast issues should be happening. Opnsense is set to let the 2 devices talk to each other, so no firewall blocking is happening. I had a rule that makes sure all DNS requests go to my unbound instance, I disabled it to see if maybe the Google display needed to talk with Google DNS servers, still same issues.

My dashboard will cast, but it is unusable. The casting only lasts for 1-6 minutes before it is recast again. Eventually the underlying software (catt) fails and won’t even try casting (after some hours of errors and partial casting).

So, does anyone else use this integration and it works for them? Just want to rule out some obvious things before I start ripping into the code and debugging.

After some more surface level debugging, I am even more confused.

I used Hass to cast a dashboard (not the same one as ccd) and it worked flawlessly. The dashboard that Hass casted is a panel type (1 card) using a webrtc card with 2 way audio. Hass does have a known issue with casting: if you try and recast to the device while a dashboard is already being cast, it throws an Invalid Namespace exception and requires a restart to get things working again. The workaround for this, is to send a media_player.stop and then cast whatever dashboard you want. So Hass has some issues with casting as well.

When I used Hass to cast the dashboard that ccd is set up to cast, it lagged and had a different issue that caused a no cast to happen (haven’t dug into that yet). Basically, the dashboard is extremely slow to load and when it does load, the lag makes it unusable. Hass casting eventually fails, just in a different, weird way. Only restarting Hass, fixes the casting issues.

The laggy dashboard is a sections dash with 2 sections and 3 cards. 2 of the cards are for lights and the 3rd card is a webrtc card. It also uses mushroom template badges for 5-6 badges. The light cards are using mushroom light entities, card-mod, expander-card in a vertical stack. The webrtc card is just the card, no extras. From reading other users posts about casting issues, I stumbled upon someone saying that using mushroom cards caused their dash to be laggy. They also said that the theme they were using (Metro?) also caused a noticeable difference. I will be testing that today after work, but from the result of only casting 1 webrtc card, that seems to coincide with that observation.

Here is an issue I have opened for any future readers or anyone having issues: Recasting: AsyncioTimeoutError / HA dashboard is NOT streaming · Issue #82 · b0mbays/continuously_casting_dashboards · GitHub