SLO monitoring and error budget

Design an SLO monitoring and error-budget system for a payments API with a 99.95% availability and a 'p99 latency < 300ms' SLO measured over a rolling 28-day window. The platform serves 40k req/s. You need to compute compliance, show remaining error budget, and page on-call only when the budget is burning fast enough to matter, not on every blip. Design the measurement, the budget computation, and the alerting policy.