What is an Error Budget?
An error budget, told as a story about a bundle of tickets you're allowed to spend on failure.
Nothing changes, so the park just stands still.Too scared of breaking it, so nobody tries anything.
Why is this hard?Scare everyone enough, and nobody tries anything new.
The slack in the target is a ticket you're allowed to spend on failure.As long as tickets remain, go ahead and try the new ride.
How much budget is left changes how bold you can be.Plenty left means go bold; little left means go careful.
Once it's spent, a rule has to make everyone stop and fix.Otherwise, guests walk away.
An error budget is the ticket for allowed failure left over by the target (SLO). While it lasts, try bold new things; once it's spent, stopping to fix comes first.
If the SLO is 99.9%, the remaining 0.1% is the budget. A burn rate tracks how fast it's spent, and teams often set a policy to pause new releases once it hits zero — a numeric balance between innovation and stability.
When grown-ups say it
- Error budget
- The slack from the target that's allowed to be spent. It's 100% minus the target.
- Burn rate
- How fast the budget gets used up. A fast burn triggers a warning.
- Relation to SLO
- The target sets the size of the budget. A tighter target means a smaller budget. → the target line set internally
- Innovation vs stability
- The tug-of-war between trying new things and staying stable. The budget turns that balance into a number.
- Budget policy
- The rule that stops things once it's spent. It pauses new feature releases for a while.
- Monthly reset
- The cycle on which the budget refills. Usually every month or every quarter.
- Burn-rate alert
- Fires when it's being spent too fast. It pages someone right away. → how many rings mean a real emergency
- SLI
- The number that measures how much is left. Calculated from the actual daily record.