Case Study
New Relic Observability Toolkit
Observability should be version-controlled, reviewable, and repeatable. This toolkit gives any SRE team a defensible, production-grade New Relic stack on day one, without spending weeks rediscovering the same configuration pitfalls everyone else already solved.
Snapshot
The Problem
Setting up observability from scratch is repetitive work. Every new environment needs the same alerts, the same dashboards, the same notification channels, built manually, configured inconsistently, and documented poorly. Teams end up with alert policies that nobody trusts and war room dashboards that mean something different to every engineer who built them.
What I Built
A production-ready Terraform toolkit that deploys a complete New Relic monitoring stack in a single `terraform apply`. Four modules, wired together out of the box: Golden Signals alerts with SRE-grounded thresholds, multi-channel notifications in SBAR format, external availability monitoring via Synthetics, and pre-built dashboards for NOC war rooms and engineer deep-dives. The defaults are opinionated because they have to be. Error rate filters HTTP 5xx only because a 404 is not an incident. Notifications follow SBAR format because actionable alerts are a core principle from the Google SRE Workbook. Every decision is documented so the next engineer who touches it understands not just what was built, but why.