Hi!

Is your feature request related to a problem? Please describe.
We have some long running outages, sometimes more than a day. An example of a catastrophic, week-long outage is this:

https://status.torproject.org/issues/2023-11-08-tor-weather-outage/

That says: "Resolved after 300h 5m of downtime". Now people strong in arithmetic will of course figure out that this is a little over a week and somewhat under two.

Describe the solution you'd like

Once we cross the 24h threshold, we should display intervals with days. For example, "300h 5m" should display as "12 days,10h 5m" here.

I guess that once you get past the 365 days mark, you might want to count in years too, but i am not sure anyone will reach that threshold.

But considering we actually use cstate to report on long-running "experiments" as well, that might actually really happen too! That's a local divergence from cstate that i'd like us to contribute back though, see our internal issue about this for more details and this component for an example. Update: opened #288 to discuss this.

Describe alternatives you've considered
I've long considered just not filing this bug and accepting the status quo, but I figured I might not be the only one finding this odd. :)

Additional context

Here's a screenshot of the outage report: