gauravs19/cloud-native-observability

Cloud-Native Observability is a reference catalog that lists what to measure and how to alert in a modern cloud-based software system. It is not a tool you run, it is a document you read and adapt. The catalog covers 316 individual metrics across 41 categories, spanning everything from frontend pages and API gateways to databases, Kubernetes clusters, AI models, and business KPIs. Metric names follow industry-standard OpenTelemetry and Prometheus naming conventions, so they translate directly into most monitoring setups. The catalog is organized around three well-known frameworks for deciding which metrics matter. RED measures request rate, error rate, and response duration. USE measures utilization, saturation, and errors for resources like CPU and memory. Golden Signals is a simpler set of four key indicators developed for large-scale systems. Each metric in the catalog has a label showing which framework it belongs to, plus a recommended alert action: page a human immediately, create a ticket for follow-up, or just display on a dashboard without alerting. The repository also includes ready-to-import Grafana dashboard files and Prometheus alert rule files, so you can get a working baseline quickly rather than building from scratch. A cloud service map in the catalog links each category to the matching services on Azure, AWS, and Google Cloud, so you can find the right monitoring source for whatever cloud you use. A companion project called NFR Advisor helps you decide which metrics from this catalog apply to a specific system, generating service level objectives for each requirement and linking them back to the relevant metrics here. Together they form a starting checklist for teams setting up observability for the first time or auditing an existing monitoring setup.