To put a finer point on it:
• Lots of working-class folks live in 🟢🏘️
• There are lots of jobs in 🟡🏭
• We’ve already built 🔵🚆
• But we’re missing 🔴🛤️
• So those workers have to drive on ⚫️🛣️
• Therefore we should #BuildTheTunnel 🔴🚇

Tbf I plan all of my rides based on what food is at the turnaround point.

Been meaning to do a tour-de-tavern-style-pizza again.

How's the breathability on these? I've been eyeing them for a while (and squinting :) )

I'd also recommend looking at these metrics broken down by user/project, and try to make sure your 1% least reliable subset is still doing ok, or at least getting support, since failures are often not evenly distributed.

I really like this post on the topic: rachelbythebay.com/w/2019/07/15...

That mirrors my approach tbh, but I'd be interested in what you write up!
I'm running a large scale testbed, and we tend to hit the worst cases from both cloud and HPC (users can request fairly weird/specific things).

We're primarily looking at "time to submit job", and its success rate.

Initially caught it because the mini-circuits label stood out from a distance :)