Loading…
A short voice session: one question, one follow-up, then instant feedback — up to 5 minutes.
Burn rate relates current bad-event frequency to the SLO's allowed bad fraction. It helps prioritize failures that threaten the reliability target.
A burn rate of one consumes budget at the rate consistent with the target over the full window. Combine short and long alert windows to detect sustained fast consumption while reducing transient noise. Use different urgency for rapid and slow burns, and handle sparse traffic and missing data deliberately. Alerts need an actionable owner and recovery information, not only a formula.