I’m definitely not an advanced user, even though I am able to follow some steps and with the help of ChatGPT and Claude, sometimes I can get to where I want to go.
That being said, when looking at this topic about having a page that’s shown when we are doing something that makes the forum show an error, a few genuine questions come to mind:
Why doesn’t Discourse have this built in? Can someone explain in plain English (as much as possible, of course, because I understand this is very technical)?
If it’s not possible now, will it ever be possible and implemented, so non-advanced users can rely on that kind of page?
Since installing Discourse 3 months ago, I noticed that 99% of the time when I try to update it directly from the Admin page, it fails and I have to rely on Terminal. I believe I read somewhere that when we update via the Admin area itself, users may experience some lagging, but it doesn’t completely go offline, the way it does with Terminal? If so, why is that, and why do I experience those errors all the time that force me to rely on Terminal?
Is there a way to update via Terminal without making it go offline? Because it seems that when adding/removing plugins, for example, I’m forced to?
Because it is too complex to justify maintaining as the free self-hosting solution and too complex to support for the average inexperienced admin?
But if you follow that guide perhaps that approach still works.
Can’t comment on CDCK priorities/roadmap
I do find it (that is “docker manager” updates) unreliable: I often find if there is an error during an “online” build, I’m unable to retry the online update after fixing the issue (usually with a plugin). The online update process appears not to respond, so I’m forced to perform a command line update. I’d raise a Support Topic if you come across an example when that happens.
Not with the basic install because that’s rebuilding a single container and because that potentially requires a new image and potentially affects the database, the whole site has to be taken offline.
There is a two container solution you could try which allows you to rebuild the web container separately that may give you more uptime: