Our management hosts have never been systematically rebooted. Some have been up for months — running vulnerable kernel versions the entire time. A security exception gave us about a year to fix it, and the interim plan was a manual weekly reboot process (#7655) — open a spreadsheet, pick some hosts, SSH in, cross your fingers, reboot, wait, check if they come back, file an issue if they don't. Repeat. For hundreds of hosts. Every week.
We wanted to kill that toil. So for the hackathon, we built it.
Imagine a grid of colored squares — each one a management host. Green means healthy, orange means it's been up too long and needs a reboot. You click "start" on a Temporal schedule, and watch.