Riemann

Riemann is a monitoring system for distributed infrastructure that aggregates events from your servers and applications and processes them with a stream processing language. It is aimed at operations staff and engineers who need to understand failures and performance bottlenecks in production systems made of many moving parts.

Streams are just Clojure functions that accept an event, and events are structs with common fields such as host and service. Dozens of built-in streams filter, alter and combine events, and you can write your own. Because the configuration is itself a Clojure program, it is concise and extendable. Examples from the site include emailing on every application exception, tracking latency distributions, ranking top processes by memory and CPU, combining statistics across a cluster and forwarding to Graphite. Riemann provides low-latency, transient shared state, and a small web dashboard shows system status at a glance.

Riemann is licensed under EPL-1.0, is distributed as downloadable packages, and has clients for various languages. You run and operate it on your own infrastructure.