Capability 13 / 18
Observability
Seeing inside distributed systems before small issues become outages.
In a distributed system, the failure is rarely where the alert fires — it's three services upstream. Observability is the discipline of being able to answer “what's actually happening right now” without guessing, using distributed tracing, structured logging, and metrics that were designed in, not bolted on after the first bad outage.
Built on OpenTelemetry-based instrumentation and New Relic, this is about catching a regression while it's still small — not discovering it from a customer support ticket.
What this looks like in practice
- •Distributed tracing and structured logging designed in from the start, not added post-incident
- •OpenTelemetry-based observability across services, not just a single dashboard
- •Performance profiling and regression hunting as a routine practice
- •Alerting tuned to signal, not noise — fewer pages, more of them meaningful