Developer & Tech

Error Budget Calculator

What this does

Track how much of your SLO-derived error budget remains after recorded downtime, with a healthy-or-breached status.

Enter your details

Runs in your browser

Calculator inputs

How to use it

Using the error budget calculator

  1. 01

    Set the SLO and window

    Match whatever your organization committed to; commonly 99.9% over 28 or 30 days.

  2. 02

    Log accumulated downtime

    Sum incident minutes from monitoring; include partial degradations if your policy says so.

  3. 03

    Act on the status row

    Healthy budgets buy feature velocity; breached budgets should freeze risky releases until reliability recovers.

Good to know

Why budgets beat binary up/down thinking

A hard 100%-uptime goal makes every blip a crisis and discourages deployment. An explicit budget quantifies acceptable failure, letting teams trade release speed against reliability deliberately instead of politically.

When the budget empties

Standard practice: pause non-critical launches, prioritize reliability work, and add page-level alerts on budget burn rate. The budget refills automatically as the old window slides away; recovery is built into the math.

How it's calculated

The math behind this calculator

budget = (1 − SLO) × window_days × 1440 remaining = budget − used burn = used ÷ budget

The error budget is the unreliability your SLO tolerates: the complement of availability spread across the rolling window in minutes. Subtracting consumed downtime yields what is left; crossing zero means every further failure breaks the SLO until the window rolls forward.

Assumptions & limitations

  • Google SRE Workbook definition of error budget.
  • Downtime measured in wall-clock minutes within the window.
  • Window assumed 30 days unless overridden.

Worked example

At 99.9% over 30 days the total budget is 43.2 minutes. Ten minutes of incidents leaves 33.2 minutes; about 23% burned, status healthy.

FAQ

Frequently asked questions

What if my downtime already exceeds the budget?
Remaining goes negative and status flips to breached; the honest signal that SLO commitments are currently unmet.
Should failed requests or minutes count?
Either works; request-weighted budgets are fairer at scale. Keep units consistent between budget and observed errors.
Who owns the budget?
Typically the service team, with product deciding how to spend velocity and SRE arbitrating when burn accelerates.

Keep exploring

Related calculators