Two's Complement Observable Metrics
8 snips
Apr 10, 2025 Dive into a lively discussion on performance-critical code as the hosts tackle the complexities of testing metrics. They debate the merits of push vs. pull systems and share cautionary tales from financial trading. The importance of structured logging takes center stage, promoting better observability and functionality. Strategies for effective error handling are explored, emphasizing communication and system integrity. Get insights into the challenges of metrics collection and learn how to navigate logging to ensure stability in programming.
AI Snips
Chapters
Transcript
Episode notes
Testing Metrics in Performance Code
- Ben enjoys writing tests for metrics to ensure correctness.
- Matt highlights challenges with testing internal cache state without altering interfaces.
Make Code Observable
- Write code that is observable for both testing and operation.
- Use metrics as a side-channel to expose internal behaviors without harming performance or design.
Use Structured Logging for Tests
- Use structured logging to capture richer, testable log information.
- Avoid brittle tests by relying on structured logs to verify internal state without direct coupling.