Error Budget

An error budget connects a reliability target to operational decisions. Suppose an appointment-booking service sets a 99.9% availability SLO for a four-week period. Its error budget is 0.1% of the measured service opportunities during that window. Whether the organization counts failed requests or unavailable minutes depends on the SLI—the service-level indicator—and the SLO definition. An error budget is therefore not an advertising budget or a pool of money.

Product and engineering teams use the remaining budget to decide how aggressively to release changes. If outages consume too much of it, a team may temporarily prioritize bug fixes, safe rollbacks and resilience work instead of feature releases. The exact policy is a local management choice; Google's SRE handbook supplies examples, not a mandatory rule for every website.

For a small service business, a lightweight version can work: define a measurable booking-form success target, monitor failures and identify when deployment risk is unacceptable. A perfectly clean dashboard is not the goal; dependable customer-facing workflows are.