Kairo

Kairo is an open-source distributed observability platform architected natively around OpenTelemetry (OTel). It is built to capture, process, correlate, and visualize distributed traces and metrics across complex microservice architectures with minimal latency overhead.

Designed the end-to-end ingestion pipeline, OTLP-compliant collector integration, context propagation across service boundaries, and web-based trace exploration interface. Implemented backpressure handling and zero-loss ring buffers under extreme throughput spikes.

High-throughput ingestion daemon built to parse OTLP spans over gRPC and HTTP. Time-series indexing in PostgreSQL/TimescaleDB, contextual span propagation across microservice boundaries, and integration hooks for Prometheus and Grafana.

Successfully open-sourced on GitHub. Tested across distributed benchmark topologies simulating multi-tier microservices, maintaining sub-millisecond serialization latency and reliable context propagation across asynchronous boundaries.