aggregateD

aggregateD is a network daemon which listens for metrics including gauges, counters, histograms, sets and events, sent over http and sends aggregates to InfluxDB. InfluxDB is a promising, but young time series database, aggregateD is intended to bring dogstatsD like functionality to Influx.

aggregateD can accept metrics either as JSON over HTTP or in the dogstatsD format sent over UDP. Therefore, aggregateD can be deployed in the same manner as either satsD or dogstatsD. That is, it can either run on the same host as instrumented applications or it can run on a dedicated host that multiple clients communicate with.

Usage and Configuration

Usuage: ./aggregated -config aggregated.json

aggregateD requires a minimal config in order to specify the InfluxDB server and its credentials. Config can either be provided as a json file or as a yaml file. An example config is as follows:

#accept metrics via HTTP
inputJSON: true
#accept metrics via plain StatsD
inputStatsD: true
#accept metrics via DogStatsD
inputDogStatsD: true

#submit metrics to InfluxDB every 60 seconds
flushInterval: 60

#output metrics via InfluxDB
outputInfluxDB: true

#influxDB settings
influx:
  url: http://localhost:8083
  username: username
  password: pass123
  defaultDB: myDB

#write to a redis list falback if InfluxDB is unavailable
redisOnInfluxFail: true
redisOutputURL: redis:6379

aggregateD exposes two web service endpoints: /events and /metrics on port 8083 by default. aggregateD accepts json encoded metrics which take the form of:

{
	"name":      		"requests",
	"host":      		"httpd.example.com",
	"timestamp": 		1461204545	
	"type":      		"gauge",
	"value":     		67,
	"sampling":  		1,
  "secondaryValues": 	{"value1: "my.host"},
	"tags":      		{"exampleTag1": 5, "exampleTag2": "value"}
}

Similarly, events are represented in the following format:

{
  "name":           "timeout",
  "text":           "a worker thread timedout",
  "host":           "node4.example.com",
  "alerttype":      "warning",
  "priority":       "normal",
  "timestamp":      1461204545
  "aggregationKey": "worker-timeout",
  "sourceType":     "default",
  "tags":           {"exampleTag1": 5, "exampleTag2": "value"},
}