SLODrift

SLODrift is a web app (with optional CLI) that continuously checks whether your service-level objectives still reflect reality. It ingests SLO definitions and telemetry from common stacks (Prometheus, Datadog, Grafana Cloud) and flags “drift” patterns: silently changing traffic mix, new endpoints dominating latency, error budget burn that only appears in certain regions, or alert thresholds that no longer match user impact. Instead of generating more alerts, it produces weekly “SLO health” reports with concrete recommendations: which SLOs are obsolete, which need tighter/looser targets, and which alerts are noisy or missing. It also tracks changes over time so you can see when a deploy, feature flag, or dependency change started degrading reliability. This is an AI + traditional app: AI summarizes findings and proposes edits, but all recommendations are grounded in measurable data and can be reviewed before applying.