Availability and latency as users experience them, per critical journey (checkout, login, sync) — three to five targets, not a metrics museum.
100% is the wrong target; the gap between SLO and perfection is budget to spend on shipping. Budget healthy → ship fast; budget burned → reliability work wins the sprint. In writing, agreed by product.
SLO dashboards, burn-rate alerts, and a monthly review where the budget drives priorities — the ritual is small; the alignment it buys is not.
Incidents produce action items and updated runbooks, not culprits — the practice that makes the whole system honest enough to improve.
Skipping the discipline this article describes until an incident, audit, or stalled project forces it — every practice above is cheaper adopted early than retrofitted under pressure.
Let's discuss how we can help you with sre error budgets slos.