Schema-First Application Telemetry
summary
Meta builds a schema-first pipeline that defines telemetry as Thrift structs, generates strongly typed emitters, and ships compact binary events through the SDK into scribe and backend stores. A CI service performs schema-oxidation CI validation to block incompatible schema changes and a metadata store serves annotations for discovery, units, and semantic types. The system reduces typos, enforces field types at compile/runtime, shrinks payloads versus key/value maps, and enables automated cross-filtering and transformation via Thrift IDL with annotations.
AI-written summary · about
chapters (6)
- 0:00Start
- 6:24Metadata approaches
- 10:30Schema-first telemetry
- 13:22THRIFT for schema authoring
- 14:36Metadata in the schema
- 16:59Annotations in the schema
related talks
29mA Practical Blueprint for Evolving Observability Schemas
05mLightning Talk: Schema Inference
31mAccelerating GraphQL API Development with Schema Storming and Code-first Implementation
32mOpenTelemetry Semantic Conventions and How to Avoid Broken Observability
27mMove Fast, Move Correctly
13mLightning Talk: Weaving Legacy and OpenTelemetry: A Schema Strategy With Weaver
32mOpenTelemetry, databases, and machine learning
29mSchemas: The Key To Data Happiness