Set reliability with an error budget, not a wish
Explain how an SLO and error budget create a practical reliability decision rule.
The move: define reliability as a user promise with a budget. An SLO has three parts: an indicator, a target, and a window. The indicator is the user-visible thing you measure. The target is the level you want to meet. The window is the period over which you judge it. Example: 99.5 percent of checkout requests complete successfully under 800 ms over a rolling 30 days. Why not 100 percent? Google SRE argues that insisting on perfect reliability can reduce innovation and force expensive conservatism. Users usually cannot perceive the difference between some levels of reliability, while the cost of…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in