baseline-monitoring

name	baseline-monitoring
description	Index of baseline metric families for Razorpay services. Use it to determine which standard metrics must exist for HTTP, gRPC, workers, egress, outbox, Go runtime, and relevant infra components.

Baseline Observability

Defines the standard metric families agents must consider when working on service observability.

This skill covers metrics only. Logging and tracing standards are separate and should be loaded from their own skills or references when needed.

Introduction

Observability in this skill is divided into three buckets:

• Baseline metrics: compulsory metrics every service surface must emit for the work it owns. These are the standard traffic, latency, error, health, and runtime signals that should exist whether the service is being created from scratch or improved incrementally.
• Infra metrics: metrics emitted by shared platform components such as Traefik, Kubernetes, Kafka, ALB, RDS, ElastiCache, SQS, and SNS. These are still part of the standard observability picture, but they should only be loaded when the service actually depends on or is fronted by those infra layers.
• Application metrics: service-specific business or domain metrics. The agent is free to define these as needed. They are valuable, but they do not replace the compulsory baseline metrics.

Usage Rules

• Always read references/baseline-metrics.md first.
• Then load only the references relevant to the surfaces involved in the task.
• Use the surface-to-reference map below mechanically: if the service has that surface, the linked metric families must exist.
• If the task only touches gRPC, read references/grpc-metrics.md and skip references/http-metrics.md.
• If the task only touches service code, do not load references/infra/ files unless the change explicitly involves those components.