Improvements to updating Cloudron: Notify admins if more downtime needed than usual
-
I've raised this before in various threads and most recently in https://forum.cloudron.io/topic/5261/6-3-3-a-few-quirks/12?_=1624981693020 so I wanted to more formally file this feature request as I don't think it exists as an actual feature request yet.
While it's rare, there are certain versions to upgrade Cloudron that requires much more downtime than usual. And unfortunately, this catches admins off guard which can cause support tickets for the Cloudron team in the forum or a frustrating user experience as apps and such are now "down" for longer than we typically expect based on the vast majority of Cloudron updates taking almost no time at all.
In my experience, most updates require only 1-2 minutes of downtime at most, many times less than a minute. In some versions though such as 6.3 that recently was released, all of the containers needed to be rebuilt which caused about 10 minutes of downtime for me unexpectedly, and I've heard some users (likely depends on how many containers and performance of their servers) encountered closer to 30+ minutes of downtime.
Speaking for myself, when these events happen and there's no heads up to the admin, I start to panic a little bit assuming something is wrong with the upgrade when I see it not coming up nearly as quickly as usual. Thankfully it just takes looking at the Cloudron logs to see that it's recreating the containers and know that it's working still, but it'd be really nice to have a heads up when Cloudron needs to recreate containers or anything else that requires more downtime than usual.
Feature request: Notify admin users if the upgrade version requires more downtime than usual.
How that's done I'm not certain... I guess change log is the best and easiest area?
Personally I always give a "30 minute" window when I schedule maintenance so my customers are usually understanding and know to expect upwards of 30 minutes of downtime in the very worst case scenario, but it'd still be helpful so we can be more accurate for how much downtime may be needed or at least when it'll be more than usual even if it can't be estimated too accurately. And at the very least, help prevent panic attacks to the admin.
