With the 2024.10 I used webRTC over the addon and it was working over cloudflare…
Regarding “live log”:
-
What is a purpose of a “Refresh” button? Shouldn’t it be removed if a log is automatically updated?
-
Are there any plans to bring back a different coloring for warnings & errors?
before:
2024.11.1:
They are colored for me. You must be running into a cache issue because the reload button goes away when you move to the live logs. It comes back when you go back to the “condensed logs”.
The live logs will also automatically stop listening after a bit and will no longer update. When that happens, the reload button will appear. Clicking it will make it live again.
Purged a cache several times. Even dared to start MS Edge! Still no coloring & the “Refresh” button is always shown.
Where can I read about this algorithm?
If nowhere - can you a bit explain how it works?
It’s just an observation that I see. Just leave the logs running and watch it’s behavior. It’ll cut out after a bit with an error, when that happens, reload pops back up.
For anyone else using go2rtc as an addon, the upgrade didn’t affect anything and I added the option to self-host and seems to work fine.
Well, if rows in Log are still supposed to be colored - here is an issue:
As for the “Refresh” button - since there is no official info in Docs about a proper behaviour, will keep observing.
If I had to guess, I would guess that the go2rtc binary (seems to have different versions depending on the machine arch its running on) did not get installed during the container update. Inside the core container (at least in my case for HassOS), the binary should be at /bin/go2rtc
.
Then it could probably cause real issues for users who use real cameras with HA docker in Debian.
Strange that this was not observed during testing.
The live logs seem quite buggy here as well. Sometimes I can scroll to the past, sometimes not. Sometimes the search/filter is working, most of the times not.
HA Core updated to 2024.11.1 but my frontend seems to be stuck:
Should that not also be 202411*
?
This is even the case in a browser with no cache/cookies/history/etc. as it was started with a fresh and empty $HOME
.
How do I get the frontend upgraded also?
Have the same problem with my cameras, dashboard tiles are OK. But trying to open give this error:
Downgrade to 2024.10.4 = OK
Having major issues upgrading past 2024.10.4 - After upgrade to 2024.11 on HAOS/RPi 4- seeing python out of memory errors and core crashes · Issue #129994 · home-assistant/core · GitHub - I think there’s something broken in 2024.11.x, unfortunately. I don’t know how to best escalate or provide logs, etc… seems like others are having problems too.
I have them colored, but depending on the Log, apparently not all logs provide the colors.
HA log is colored, and Supervisor (well not sure, but see only green so take it all is well ) but eg Host Z-wave js, and MQTT, and many other add-ons are all black.
Talking about those logs, I now notice this in the HA log (inside the HA interface)
and have that more than once.
Any ideas what that could be?
They are only colored if the line contains INFO, DEBUG, WARNING, or ERROR
not sure about that, check DNS/ESPHome logs. INFO on all lines, and yet all are black
Like you and others, I have the same issue.
Heavy memory leak in HA core leading to crash and reboot after a few hours.
I immediately suspected the new Go2RTC implementation and as a test I restarted my Go2RTC add-on and configured the Go2RTC internal integration to point to it.
And bingo !, Now Core memory is stable but the Go2RTC add-on explodes after about a day due to memory leak.
Before core 2024.11, the Go2RTC add-on was more or less stable, so I guess the culprit could actually be the new Go2RTC internal integration making the Go2RTC server go mad.
My next test will be to get rid of this integration (by painfully manually setting all default configs) and go back to the old trusted WebRTC integration, because this new Go2RTC internal integration isn’t working good enough yet for me, it kind of works for a few hours then all my camera eventually go back to slow HLS and I have to restart HA.
So I guess, it still needs some work…
There are other stipulations that must be met related to where the LOG_LEVEL is located in the line with other special characters. This is typical of any log-coloring library.
You’ll notice every line that has [time] LOG_LEVEL: message
is colored. Also every line with date time LOG_LEVEL (thread) [path] message
is also colored. This is just how log coloring works, it’s fairly standard across coding languages.