2021.7: A new entity, trigger IDs and script debugging

No problems at all here with 2021.7.1 in Firefox (89.0.2, 64bit) on Win 10. All smooth and snappy like ever. Maybe one (or more) FF-Add-on’s are making trouble? I would try FF with all plugins disabled and if the problem goes away switch them back on one by one until the HA UI/Logbook etc. slows down again.

I doesnt found a belongig post in this thread so im writing my issue in hope it hasnt been talked through here

Since the update my samsung tv integration has a weired problem - the tv works as expected but if the tv is off it turn unavailable instead of offline. Anyone else has this problem?

100% HA that causing the slowdown especially on FF. Eventually it will make FF to a halt. On Safari however it did not do the same thing however I receive a warning.

Sorry, gotta go back to the History Graph Card.
I can live with the color change, but what’s irritating me is that I hide one of the lines in a graph and then, 15 seconds later, the line shows up again without me changing anything.
You can even see in the example below that the line is supposed to be hidden (by the strikethrough) but it shows up nevertheless:

Happens on Chrome as well as Edge - anybody else experiencing this?

1 Like

No issues here. Are you using the hub? ISY?

I can confirm this. Takes less than 10 seconds for me, think that depends on Lovelace ui refresh interval. Anyway, that´s a bug, not a feature :smiley:

What´s even a bit more irritating to me is on the line history graphs the entity name on the left is almost not readable because text is cut (blacked on the screenshot):

Now we cant set attributes in the base component there seems to be no way of setting entity_picture to an MQTT sensor.

this used to display every users avatar.

I can see why they made the change but I think it could be really handy to allow the entity picture to be overwritten.

The hub. I also updated two add-ons at the same time, but I reinstalled them both after the full restore, and still no issues. The log showed that it couldn’t connect, but I should have copied the details. I may have to do the update again to get the specifics.

Has anybody experienced HA stability issues? I have HA deployed on docker in a raspberry pi 3b (official docker image), and for the first time in years I noticed that HA is unstable, restarting itself once or twice a day, since the upgrade to 2021.7.1. I was running 2021.5 and .6 prior to this with no issue. I could not yet narrow down the component that could be causing this. I use zwave-js-mqtt, multiple cameras (stream integration) and many REST sensors that I developed myself.

I have solved it with this; homeassistant 2021.7 device automation triggers · Issue #385 · AlexxIT/XiaomiGateway3 · GitHub

1 Like

I think this may be your solution: Fix Insteon component with python 3.9 by teharris1 · Pull Request #52927 · home-assistant/core · GitHub

Looks like it will be part of 2021.7.2

Have you logged a bug report for this?

Trying to gauge first if it’s only and issue for me/my installation or if it shows up for other people as well; haven’t seen many folks saying that it’s an issue for them, too.
Do you observe the same behavior?

Same issue? Any fix / workaround for this (besides a rollback)

Well, you made me look. I have 2021.7.0 running on my test server and, after staring at this graph for a minute, I can confirm it didn’t display anything other than what I asked it to display.

Tested with MS Edge on a Windows PC.

No, not yet, as I haven’t updated yet to 2021.7 but someone reported it already.

1 Like

My first attempt – no dice. But I picked a graph with data that changes more rapidly and sure enough, everything comes back with the selections still lined out.

My guess is that the deselected trace will reappear when whatever is plotted changes from its last value.

I see it with 2021.7.1 on Win10 with Chrome, Edge, and my Android devices.

It might be that my test was insufficient. Inkblotadmirer mentioned:

My graph didn’t have anything that updates frequently so that might be why I didn’t observe the problem.

This is a frontend bug, not a core bug, FYI.
A mod there will probably move it for you if they can, or will close it and ask you to re-open it against the frontend repo.