Telemetry as Code (TaC) is the practice of defining, versioning, validating, testing, and deploying telemetry collection, processing, enrichment, routing, and export configurations through code and automated delivery pipelines.
Once telemetry has been generated, another problem starts: What should we do with it?
This is where I use the term Telemetry as Code.
I define Telemetry as Code as the practice of defining, versioning, validating, testing, and deploying how telemetry is:
Collected → Processed → Enriched → Filtered → Sampled → Routed → Exported
OpenTelemetry Collector is a great example of an enabling technology for this model.
Instead of manually maintaining Collector configurations on individual servers, these configurations can live in Git and move through a controlled CI/CD process.
With technologies such as OpenTelemetry Collector, OpAMP, Kubernetes, GitOps and CI/CD, telemetry management can become a platform capability rather than something every application team has to solve independently.
This also creates an important abstraction:
Applications generate telemetry without needing to understand every observability backend. The platform determines where and how that telemetry should be processed and delivered.