Python Developer

About the Internship

During my two-week internship at Eywo, I developed a stress-testing tool for the Jaeger distributed tracing system. My primary objective was to build a test data generator capable of identifying specific system bottlenecks and anomalous behavior under extreme load conditions.

Trace Data Generator for Jaeger

I developed a Python-based utility to generate complex trace data (spans and traces). This tool was instrumental in investigating a performance issue where the Jaeger dashboard would fail to display more than 60 concurrent traces.

• YAML-driven Configuration: The script utilizes YAML files for configuration, allowing users to define the volume of generated traces, their minimum/maximum latency, and overall duration.
• Structured Data Modeling: I implemented logic to generate JSON objects that represent spans with realistic parent-child relationships and logical sequences, ensuring the simulated data remained valid for analysis.
• HTTP Integration: The generator pushes data directly to the Jaeger ingestion service via standard HTTP requests.