Hiya Rocket Chat Community!
First off, I want to start out saying amazing work on getting this piece of software together. Based on many search attempts, I don’t think there is any good alternative open source chat system such as this. However … We have some problems
One of the biggest things that has kept me away from using Rocket Chat beyond a testing environment is the sheer amount of issues/bugs it seems to possess ultimately making it unstable for long-term production use. Yet, I am noticing the community seems to still be quite focused on Implementing new features ATM.
One thing I’d love to see happen is a release with focuses solely on stablising the current code so it’s truly production ready. Compared to alternative OSP, nothing comes close to the features which are already here, but aren’t very stable. By focusing on fixing what is here now, Rocket Chat can truly become a competitor to the likes of Slack, Microsoft Teams, Discord, etc. But as it stands, the walled gardens are the more likely choices due this huge flaw of instability.
One user suggested implementing a LTS branch, which I think would be the best way to tackle these issues while still working toward new features. I personally think this would be the best approach overall for the long run.
A LTS version of course has a certain appeal and makes sense in certain environments. The question is whether this is the case with Rocket.chat, as the project is clearly in flux and is being developed very actively. We’ve been running Rocket.Chat productively for more than two years and I can’t say that the installation has let us down even once. So I think that Rocket.Chat can be used and recommended in a productive environment with a clear conscience. Regarding possible bugs it helps to report them on github. From my own experience I can say that all bugs I posted there were fixed when a clean issue was posted. This is an agility that I really appreciate and I wouldn’t want to be nailed to a version for five years like Ubuntu Server LTS. There’s still too much going on in this segment for that. And separate maintained version branches are more time-consuming. The question is whether it’s better to bundle resources on a rolling versioning. Updates have to be installed anyway, whether LTS or not.
I fully agree with LTS idea. If you want SW running in production enviroment, you need a very stable product. So far, nearly every update of rocket chat is with expectation that something will stop work or even proces of update do not make it to finish line. You can take example from debian release process…they freeze development and adding of new sw at some point and get focus only on hunting bugs, to get to the final release to users. And at the same time, they continue from freez point as another branch and adding new feauters and sw.
Sooner or later team behind RCH will get to the point due to the growing community, where they will need stable product in order to get hands free to develop new features. I thing this will come much sooner then they think.
I’m not sure if any of you have really noticed but Rocket has been at Alpha stage really - anything less than a v1.0 release is exactly that. Current stable is 0.74.3 - but it is still really Alpha.
The fact that it HAS been so stable for such a long time is credit to the devs, particularly as they were trying different things all the time. I too have been running it without issues since around 0.20 - it really has been pretty stable for me, though I have taken my time on upgrading it.
So you are attacking them without realising what stage the software development was really at - a cursory glance at the documentation on Github would have told you, and installing and updating was at your own risk.
If you were using snap you would have also noticed that they delayed release to try and ensure some stability there as it is probably the largest deployment method. However, the devs also got criticised for NOT upgrading snaps fast enough because some users wanted new features…they can’t win.
Yes, if you are paying for a supported version then you have an avenue to complain if things break on update and will get help, but I don’t suppose that is the case. So you have got what you paid for.
The simple answer is don’t update until you have tried and fully tested a newer version, unless there is a really pressing need such as a security fix . That’s what professional, sensible, sysadmins do.
I would imagine the pace of development will slow once v1 is out the door. There will be as now a stable version, and as now a ‘bleeding edge’ development version. Probably sufficient for most people.
