eBPF-based telemetry agent that collects and processes systems telemetry. When fused with application metrics, it enables business-to-metal analysis of any runtime. Recent versions ship with a built-in chart viewer and an MCP server.

Benchmarking tool for LLM inference services with detailed analysis of throughput, latency, and resource utilization across models and serving configurations.

Performance measurement tool for popular RPC services including Redis, Kafka, Memcached, and gRPC. Highlights the relationship between throughput and latency distribution with detailed histograms.

Extremely efficient library for static metrics. Supports all metric types with an ergonomic, distributed declaration syntax.

h2histogram, a base-2 HdrHistogram designed for the hot path. Cheap to update, pre-allocates memory, and offers fine-grained precision control.

Precise and efficient access to system clocks for performance-sensitive applications.

Token generation library for rate control in high-throughput, latency-sensitive scenarios such as load generation, traffic shaping, and latency measurement under precise request rates.

The most popular Rust wrapper of the NVIDIA Management Library. Originally by Jarek Samic; maintained by IOP Systems since 2024.