WTH - Why does it seem like lots of github issues are auto closed

You have 7 days to reply.

1 Like

FWIW, I have posted several Issues and they fell into three categories:

  1. Received reply and were eventually fixed.
  2. Received reply but were never fixed.
  3. Never received a reply.

For categories 2 and 3, I continued to comment (“Issue still present in current version”) whenever the system announced the Issue was stale and would be auto-closed. I recall doing this for one particular Issue for two years before allowing it to auto-close, unresolved.

I have resigned myself to the fact that there are only so many developers willing to contribute their time and, when they do, they spend it on things that interest them (and not all Issues interest them).

1 Like

It’s almost like they are… volunteering their time :thinking:

1 Like


The same appears to be true for Nabu Casa’s full-time developers (whose salaries are funded by payments/contributions from users of Home Assistant).

All this to say, I’m in no position to change the system so I just accept it for what it is and use the free software I get (bugs and all).

1 Like

While I agree with your comments, I think where the community expectations might differ is when there are paid full time developers backing the highly successful and active project (which greatly differs from oss developers work on what they want to work on). I pay to support the project. I don’t expect to get anything from it, and I don’t think people are asking for support. With this said I do think they/we all expect real issues to be addressed even if it’s in a verified bug backlog if there are paid resources (I get this contradicts we don’t expect anything from it). It would be nice if there was a full-time project manager / person prioritizing / dev advocate. Again, this is just my feedback/observations. Thank you for everything you do.


There is a person who does it, and other devs help out from time to time. Just because there isn’t a ‘clear’ priority marked in issues doesn’t mean it’s completely ignored. There’s an entire tag system in place to help with all of this. There are tags to fast track things, there are tags to let the devs know the current state. A lot is being done and managed, that’s what’s funny about all these armchair project manager comments here. Everyone thinks they have a full proof method until they get into the fray. I’ve been in the fray, I can tell you that it works even though it’s a bit slow moving.


this has got to be the most abused phrase in HA replies to disgruntled complainers.

Honestly, HA has come of age, at least so much it should pride itself not to use those words again. It has become so professional and has such a large team of pro devs currently, the HA management should at least refrain from the ‘volunteer’ stance.

Ha/Nabu Casa isn’t free anymore. I costs more than any of my other software online services (PIA, Tunnelbear, cloud storage etc etc, Plex).

Let’s stop talking about volunteer software please.

Doesn’t take away from the prioritizing scheme, I am fully aware of that. And like Taras, try to go with the flow.
Some times that is particularly difficult though, like in the constant logout (and auto-bans) issues we experience widely. Just an example. Sure there are others.


It doesn’t change the truth behind it.

No, this is an open source software. It 100% relies on volunteers contributing to the codebase. This is what keeps the system alive. Without volunteers, we wouldn’t have a large portion of the integrations.

Paying for Nabu Casa gives you secure external access to your system through a paid service.


They are volunteers though. Nabu Casa is offered as paid but HA is free with an Apache 2.0 license. That means it comes with no warranty and is offered as is. There is no guarantee of functionality or support.

HA status as a viable option is irrelevant to the fact that it is open source software available for free and supported by passionate devs in their free time.

1 Like

no, this can not be maintained. I know that is what it started out like, but the team does way more than that. And they’re all paid by the subscription fees. NC is the professional company running the Home Assistant project. And I am glad that is what its growing into, don’t get me wrong. They are all wildly proficient in what they do, I have the utmost respect for their capabilities.

It’s is not a project of volunteers however. It might rely on many volunteers spending their time to test, write integrations, write-up issues, install testing CC’s for the devs to check the functionality, help out others in need (like yourself thank you very much). etc etc.

I stop here because it diverges from the original auto-closed WTH.

It would be good if the final week span of closure could be widened somewhat, so we testers have time to test and respond to it, while maintaining another life too…

1 Like

It’s not like your next reply has to be if the issue is solved or not. Can easily just reply “will check this weekend” to make the stale bot go away. Then after checking write another comment or close the issue manually.

Honestly, I think 7 days is pretty generous, over at Frigate we have stalebot set to mark stale after 30 days and then close 3 days later. It is aggressive because a number of issues people won’t respond to for some time especially if it was fixed and it adds a lot of clutter otherwise, difficult to see what is actually still a valid issue.

well yeah, that could be a way of responding too I guess. maybe its not such a big fuss with that approach.
Btw, I see I forgot to mention Ha is also more expensive than my 1Password subscription for 6 people… and that at least has Touch-id/Face recognition :wink:
O, and never logs me out so I need to re-install the App allover… :grimacing:

That’s not HA though, that’s nabu casa (external access to HA). I have a cloudflare tunnel setup which adds external access to HA along with 5 other services on my server and besides the $4 to renew the domain each year it is entirely free. I support HA through my contributions to core and the HA Android app (along with the addons I contribute to).

I understand not everyone has the want to spin something up themselves and also see Nabu Casa as a good way to support the project. However, conflating the cost for Nabu Casa and HA is not a true representation of what is being paid for.

1 Like

I think a lot of people here don’t get how many people file an issue and NEVER come back and respond. It happens in every repo and every project. If you are here and you noticed it then that means you pay more attention than most users do. There are a lot of drive by PRs and one time issues that get filed with no follow up at all from the user. The simple truth is if you have an issue then you should be responding frequently to the issue. Either providing more information or simply retesting the issue in each update.


No. there is no such obligation.

you require users to be testers or do the another job for this project. Don’t know where this confidence you took from. if you think that anyone who uses open source software is obligued to contribute to it then you are pure wrong.

Most users want to be just users. They don’t even report issues.

If some one reports an issue it should be considered as contribution. so isn’t it what you want from users?

And if a bot throws this effort to the bin, how it should be called? waste of resources?
For me personally, closing issue without even taking a look on that is sign of disrespect.

and as already said: I never experienced this in any other project. There are issues prioritized in way that they wait long time if I or someone else push it better. But were never closed the way HA bot does

No but its the right thing to do, out of the thousands of issues that get opened its good for users to want to do more and help move the project forward. Opening an issue and never responding doesn’t help anyone, even the user who opened the issue.

this is where the “community” term comes into play, everyone including users like yourself are part of the community. Whether you participate in the forum or submit a PR you are part of that community. You don’t have to report an issue but if you want it to get looked at and taken seriously then yes you should do more.

probably 90% of users or more do this today and may not even know this forum section exists

As the project moves fast and lots of code changes come in its perfectly reasonable for issues to get resolved by other fixes that were unintended. So as the user experiencing the issue it helps to retest and report back if the issue is the same or if the behavior changed. The more data you provide the more likely it is to get fixed so why not provide as much data as possible? Don’t you want the issue resolved?

That is the reason why the bot says you have “x amount of days to respond”, it takes a minute of your day to respond to an issue. How long do you think it takes a volunteer to create a fix, test it, submit it and get it reviewed? So its perfectly ok to expect volunteers to spend lots of free time to fix issues and add features but its disrespectful to expect a user to respond to a bug report in a timely manner?

how many issues do those projects have? Does the issue list seem like something that can be managed by say a person who needs to review the entire list? You have to trim the edges and get the bad reports out. How many issues do you see about users needing help and not of actual bugs? How many feature requests get reported as bugs? How many issues are actually fixed but still remain in the issue tracker?


I really try to close any issues I opened and can confirm that they are fixed. I feel that this is the least that I can do. I really appreciate the efforts from all that are volunteering their time to this great project.

I am however frustrated by some of the issues that have been on going for years that I only see rarely. A good example is in a system restore. The restore UI is broken. It is very confusing at the least (at least it was) but I only test this when I have a system issue and have to do a restore. This maybe (hopefully) a year or more from now. Related to this was an issue on the Mariadb not starting on a restore. I received a bot yesterday that this is now stale. I have no idea if this is fixed ( or even looked at) and I have no way of testing this prior to it being closed.

If someone has looked at any of the open issues it would be helpful if a message is sent to the telling the issuer to please confirm that it is still an issue.

So just leave a comment saying that you plan to test but need more time to do so, or just leave a comment saying “ping” to give yourself more time. Again, it’s not like you have that many days and your only response can be “fixed” or “not fixed”

They do, but again as was said previously unrelated fixes may touch that part of the code and subsequently fix the issue you saw as well so whoever wrote that did not know to update the issue. Or someone fixed it without realizing there was another open issue for it so they did not leave a comment.

it’s a different case. I gladelly provide more information, logs, screenshots, try to reproduce the issue differently to the developer’s response.

What we are talking about is closing issues which wasn’t even looked by devs at.
And I don’t want to play the game with the bot. This is wasting my time giving no any returned value (because there is no guarantee that repeating the refusal of closing the issue leads to its resolving).
It’s not about “how many time I need to respond”. It’s about the fact it’s an anti-pattern. Effort which leads to nowhere. Besides such a detail that with the current release cycle probability that issue is incidentally fixed by another task during a month is very very low. And if it happens it means that nobody takes care about reported issues to link them as duplicates.

This leads to the final question: why should I care to report anything?

You might answer: go away, we don’t want you. But once you say that to all those valuable reporters who will stay to help the project , who will remain?

You don’t know if it was or wasn’t looked at. A dev could have tried to reproduce or looked into it and felt they were unable to help so left it alone (I know I have done this at times with the projects that I contribute to).

Also many times users incorrectly report issues which leads to no code owners being assigned so it goes unnoticed (since appropriate person was not notified). This is the due diligence of the reporter to read the steps and report the issue correctly (which often does not happen). Of course there are people that go through and fix that, but there’s 1000+ issues in the core backlog alone so it is quite the job even do that.

It seems users often look at their issue in a vacuum. It is affecting them so they naturally find it more important than other things that are worked on, but volunteers and others contributing only have so much time. There are hundreds of issues created every week on core alone, along with a 1000+ backlog. To say it is unreasonable for an issue to get accidentally missed or not be addressed within 28 days is disingenuous.