Sloth

"Sloth: Generating SLOs for Prometheus Without YAML Nightmares"
If you already run Prometheus at scale and you’re tired of brittle, copy‑pasted SLO rules that drift across teams, this book is for you. It’s written for experienced SREs, platform engineers, and observability leads who want SLOs to be an auditable reliability contract—not a pile of hand-tuned PromQL and unreadable YAML diffs. You’ll treat Sloth not as “magic,” but as a disciplined compiler for reliability intent.
You’ll master the Prometheus mechanics that make generated SLOs trustworthy: robust PromQL SLI patterns, recording-rule design, rule-group dependencies, and alert evaluation semantics at the Alertmanager boundary. From there, the book dives deep into authoring Sloth’s prometheus/v1 spec: service identity and label strategy, defensible objectives and windows, correct event-based queries, and alert ergonomics that support routing and runbooks. A full chapter on multi-window, multi-burn-rate alerting turns error budgets into actionable page vs ticket signals, with tuning guidance for noise, low traffic, and missing data.
You’ll also learn production workflows—validation gates, deterministic generation, CI/GitOps delivery models, and operational verification—plus Kubernetes-first patterns with Prometheus Operator and the Sloth controller/CRD. Finally, advanced chapters show how to debug alerts back to spec intent, extend Sloth safely with plugins, and manage version milestones so your SLO platform stays maintainable over time.