Well, it’s always frustrating when there is unplanned server downtime.
Today we upgraded the operating server operating system, and most of the server stack as well. This was expected to be a relatively brief outage after making backups of everything. Unfortunately the system upgrade took more than two hours, and another stretch of time for the 800+ packages involved. There are still a couple of sites offline awaiting non-essential manual over-rides.
For which I am deeply apologetic. I plan to address this over the next couple of weeks, developing an alternative model for system upgrades.