Can we stop with the uptime percentages?

This argues that uptime percentage figures on service status pages are a misleading way to communicate reliability to most users. Citing the idea that each step from 90% to 99% to 99.9% is as hard as the last, the writer describes increasing daily encounters with outages (GitHub, CI, AI tools, Slack) and the confusing wall of near‑100% numbers. Those percentages hide nonlinear differences: tiny changes near 100% can mean an order of magnitude difference in downtime, the kind of distinction infrastructure people intuitively get with “two nines” or “three nines,” but normal users do not.

The practical proposal is to show absolute downtime alongside or instead of percentages so anyone can grasp impact without mentally translating magnitudes. For example, instead of “GitHub Actions: 98.31% uptime,” display “GitHub Actions: 12 hours affected in the last 30 days (98.31% uptime).” Presenting hours affected answers the straightforward question users actually have - how much time was lost - rather than expecting them to parse percentiles that look similar but imply very different service behavior.