Picture courtesy prosperwithit.com
It had to happen: After a carefully planned move of the Dailykanban servers, everything was good to go by Sunday evening. Except by Monday morning, all servers had rolled over and were (pretty much) dead.
Dailykanban doesn’t just run from one server, it follows the belt-and-suspender principle, and runs from a bunch of servers, in all corners of the world. They talk to each other, and keep themselves entertained and in sync. In theory, at least.
Apparently, the master server was dinged during the move. All other servers happily updated from the master server, and contracted the same ding thing.
We fared quite well in the past with the belt-and-suspender principle. In the more than 10 years of Dailykanban history, we never were down more than the 10 seconds it takes to switch from one server to the other. Until today …
And why did we move at all? For more than 10 years, we rented our servers from a Munich company called Contabo. Us Bavarians need to stick together, and their low rates fit our non-profit venture perfectly. Except that the big enshittification has caught up with Contabo. All our servers suddenly got unbearably slow. A postwar teletype was faster than the server console. You could hear the bits drop one by one. Contabo had been bought by a private equity firm, and the private equity firm needs to get its money back. So they apparently massively oversold their servers. Not only were the servers overloaded, their “support” people were too, and started lying through their teeth. It was like owning a Tesla: Everything was “in spec.” After two weeks of battling with Contabo, we called it quits.