What?
A stream processing language and compiler for human-scale infrastructure monitoring
"The right solution for 100X often not optimal for X" — Dean Jeff
Why?
Those last years, thanks to such large companies as Google, Facebook, Linkedin and Netflix, the culture and practice of modern infrastructure monitoring has vastly improved and many good and free tools have been released publicly. Those tools understandably focus on large distributed infrastructure.
For smaller use cases though, tools have been left to where they were in the 90s, with the notable exception of Riemann. But Riemann is only for monitoring hosts and uses Clojure as a configuration language. Which in turns requires a resource hungry JVM.
If you need an all-purpose stream processor to manipulate time series in order to turn inputs from sensors or network probes into alerts but do not want to deploy Kubernetes in your three racks of hardware or have only a couple of GiB left of RAM for monitoring, then you might want to consider Ramen.
How?
This is how operations look like:
DEFINE memory_alert AS
FROM memory
SELECT
time, host,
free + used + cached + buffered + slab AS total,
free * 100 / total AS used_ratio,
used_ratio > 50 AS firing
GROUP BY host
COMMIT AND KEEP ALL WHEN COALESCE (out.firing <> previous.firing, false)
NOTIFY "http://192.168.1.1/notify?title=RAM%20is%20low%20on%20${host}&time=${time}&text=Memory%20on%20${host}%20is%20filled%20up%20to%20${used_ratio}%25";Currently the stream processing programs are compiled into a language with automatic memory management (OCaml), so performance is not optimal. The plan is to compile down to C (or such) in a future step.
Also, imports/exports are limited: Ramen currently accepts time series from CSV files, and understands collectd and netflow (v5) protocols. As output, it merely reach out to alerting systems via HTTP requests.
Other than that it is possible to “tail” the output of operations from the CLI. More protocols for both input and output need obviously to be added.
-
Docker image: rixed/ramen:demo on https://hub.docker.com