I have. Your WTH is just not possible and you are bitter of people telling you so.
Home Assistant is open source for many, many users. Just that you don’t want to fix the issue with 15 errors per second, does not mean the whole of Home Assistant has to change because of you. The logs are there for people to debug or just for reference and information. You can already edit, adjust and adopt all the items you want to appear there. The summary is already there in the UI for people who don’t like using code.
That you are too lazy or ignorant to change a single line of code or select an option in settings is on you. No need to piss people off or call names.
no system should hide errors like this. As an Architect level system admin I would be furious if a system starts hiding error messages… Yi would consider the race condition an issue itself and yes it’s because something needs to be addressed.
You may disagree that’s fine. But in real IT land that how it works and you will likely not get support for it.
But were beyond that. How about everyone step away from the keys for a few minutes K?
If they occur in such a short time period yes. It points to a larger issue with the configuration or integration. Nothing should be polling that fast (16 times per second).
That’s pretty silly. You need 200k of the exact same, full detail, message posted with the only change being the timestamp to know it’s a “larger issue”? I sure don’t. But maybe I’m just far ahead of most.
Everybody is so afraid that what I’ve suggested might take away something from them. It’s quite…silly.
Chill. There is no need to get upset. Nothing anyone says in response here matters. It is the developers reviewing the WTH that will decide the worth of your submission.
As a dev let me tell you that the reason I oppose this change is precisely because I care strongly about errors. The notion that some logs should be just dropped, because you have misconfigured your instance, is a terrible idea.
Maybe a constructive way forward here would be a suggisting some kind of feature where you are notified in the UI if a error log message exceeds a certain count? Might be hard to come up with a meaningful filter that covers all bases though.
if that number increases rapidly, I know I have to check out the logs…
and fix it.
is exactly this
large log files will not crash HA , it’s the underlying problem that will.
that’s why the summary you show is an important red flag, summoning you to check what’s up, and next fix the error.
how can one oppose that system
most of the times these frequent errors stem from an automation (had a famous killer myself) or templates.
Both of them need to be tracked down, and in an ideal world I suppose the system could disable those.
I seem to recall a mechanism was added to automations to limit too frequent and unlimited firing, though not sure about that. Would be cool.
making the WTH into 'Why is there no auto-defend against frequent offenders"
Beg your pardon, but I’ve had it happen. HA is parsing the log file, no? Which is just a text file…non indexed, nothing special. There’s a limit to what it can handle.
Auto-disabling a frequently failing automation, template, or integration would be great as well. But I know that’s not going to happen.
