Explain the differences between SLI, SLO, and SLA with concrete examples for a web API (e.g., availability and latency). Given a quarterly target SLO, define a reasonable error budget and show how you would apportion and track its consumption over time. Then design a monitoring and alerting system that minimizes alert fatigue: choose which signals to alert on, set thresholds relative to SLOs, aggregate and deduplicate alerts, apply multi-window/multi-burn-rate policies, and define escalation, silencing, and runbook practices.