Metrics Stream Processing Using Riemann (EN)
This talk will cover:
Introduction to Riemann
* In-memory stream processing system written in clojure
* How it can easily process millions of events per second
* Configuration is clojure code
What problems Riemann solves in the today’s world of Prometheus & Datadog
* High cardinality metrics problem: Since all processing is done in*memory, riemann can handle high cardinality issue much better than prometheus.
* Instant detection: Since riemann uses websockets, an issue is instantly reflected into dashboard. In prometheus, it will only get reflected after next scrape.
Riemann concepts
* Event
* Stream
* Index
* Integrations
Riemann Stream Processing Engine
* Types of functions on riemann streams.
* Examples:
* Combine multiple streams into one stream.
* Split one stream into multiple streams.
* Filter, roll up, throttle, coalesce events.
Extending Riemann
* How riemann schema is extendable and use for streaming other events like logs.
Demo
* Using docker-compose setup of riemann to present the power of riemann and riemann-dash.
Speaker
-
Pradeep ChhetriCloudflarePradeep is an SRE at Cloudflare. Originally from Nepal, he is now based in Singapore. In his free time, he enjoys travelling and hiking.