Positional Paper: Schema-First Application Telemetry
Application telemetry refers to measurements taken from software systems to assess their performance, availability, correctness, efficiency, and other aspects useful to operators, as well as to troubleshoot them when they behave abnormally. Many modern observability platforms support dimensional models of telemetry signals where the measurements are accompanied by additional dimensions used to identify either the resources described by the telemetry or the business-specific attributes of the activities (e.g., a customer identifier). However, most of these platforms lack any semantic understanding of the data, by not capturing any metadata about telemetry, from simple aspects such as units of measure or data types (treating all dimensions as strings) to more complex concepts such as purpose policies. This limits the ability of the platforms to provide a rich user experience, especially when dealing with different telemetry assets, for example, linking an anomaly in a time series with the corresponding subset of logs or traces, which requires semantic understanding of the dimensions in the respective data sets. In this paper, we describe a schema-first approach to application telemetry that is being implemented at Meta. It allows the observability platforms to capture metadata about telemetry from the start and enables a wide range of functionalities, including compile-time input validation, multi-signal correlations and cross-filtering, and even privacy rules enforcement. We present a collection of design goals and demonstrate how schema-first approach provides better trade-offs than many of the existing solutions in the industry.