Tracking uptime percentage vs. an error budget — which number should drive decisions?

Tracking uptime percentage vs. an error budget — which number should drive decisions?

Q: My dashboard shows 99.95% uptime. Is that the metric to manage by?

A: A raw uptime percentage is fine for reporting, but it's a weak decision tool because it hides time. "99.9%" sounds great until you do the math: that's about 43 minutes of allowed downtime per month, 8.7 hours per year.

An error budget flips the percentage into a spendable allowance. If your target is 99.9%, you have ~43 minutes of "budget" each month. Burned through it on a deploy gone wrong? Freeze risky changes until the window resets. Plenty left? Ship faster.

The practical move is to track budget remaining, not just the headline percentage. It turns uptime from a vanity number into an actual lever for how aggressively you release. The percentage tells you how you did; the budget tells you what to do next.

Got a question? Drop it in the comments.