1
0
Fork 0
Re-imagined Graphite/StatsD stack ready to run in Docker
Go to file
Gregory Eremin d548ff0765 Allow CORS in graphite api 2016-08-09 14:01:29 +02:00
configs Allow CORS in graphite api 2016-08-09 14:01:29 +02:00
Brubeck.dockerfile Change Brubeck port 8080 -> 8090 2016-07-07 21:18:16 +02:00
Carbon.dockerfile Add graphite api image :( 2016-07-07 21:20:45 +02:00
Grafana.dockerfile Shorten dockerfiles (since they build fine) 2016-07-07 13:04:38 +02:00
LICENCE Make graphite fast again 2016-07-06 21:59:26 +02:00
Makefile Add build command too 2016-07-08 16:16:51 +02:00
README.md Map data directory to a host machine 2016-07-25 14:49:49 +02:00
docker-compose.yml Map data directory to a host machine 2016-07-25 14:49:49 +02:00

README.md

Graphite-ish

Graphite project consists of three components that designed to work together:

  • carbon daemon that accepts and stores metrics
  • whisper database where metrics are kept
  • web app that reads metrics and processes them using a comprehensive list of supported functions

Every component exposes several APIs that are used by other components. All of them are written in Python.

Two more components are often used together with Graphite:

  • StatsD — a Node.js app that accepts metrics and aggregates them before submitting to carbon;
  • Grafana — web dashboard for Graphite (and other backends)

So what?

This set of images replaces original carbon daemon and StatsD aggregation daemon with other compatible components to improve performance and throughput.

Grafana 3 is also included.

Included images

Component Version
go-carbon v0.7.2 Drop-in replacement for original carbon daemon, written in Go
Brubeck 5d139a4 Easy replacement for StatsD, written in C
Graphite API latest A piece of original Graphite Web component with less features
Grafana v3.1.0-beta1 Beautiful dashboard for all of that

Building

Make sure Docker and Docker Compose are installed.

$ make graphite fast

Re-building

$ make graphite fast again

Starting

$ make run

Using

  • Open Grafana dashboard at http://[docker-host]:3000/, sign in as admin/admin
  • Create new data source of type Graphite that points to http://graphite-api:8000
  • Create a new chart with a test metric (test.ping in this example)
  • Start sending metrics to StatsD-compatible endpoint
  • Metrics should show up on Grafana dashboard

You can send metrics right from terminal:

# 1000 times
# Send "test.ping:1|C\n"
# Verbose, Using UDP, Timeout 1s
# To StatsD port on Docker
$ for i in {1..1000}; do echo "test.ping:1|C" | nc -v -u -w1 `docker-machine ip` 8126; done

Data is kept in /data/graphite/whisper directory on the host machine.