Registry backends

Simple in-memory registry

This basic implementation is available in core library. Is is simple in-memory implementation that does not integrate with any specific monitoring system. It can be queried via

Creating simple registry:

(configure 
  {:type :simple})

Composite registry

Composite registry binds several registries together and presents them as single entity:

(configure 
  {:type :composite,
   :components {
     :local {:type :simple},
     :cloud {:type :prometheus}}})  

All component registries will be accessible via :components key of returned object.

AppOptics

Sends metric data to AppOptics service.

Add POM dependency: [io.micrometer/micrometer-registry-appoptics "1.5.1"]

(configure
  {:type :appoptics
   :api-token "79d84fcf-ef37-4297-9c98-9e912fed7556"
   :host-tag "my-app.svc.kubernetes.local" })

It is a push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :api-token - API token to authorize data submission requests (required);

• :host-tag - unique identifier of monitored application (required);

• :uri - data collector endpoint URI (default: https://api.appoptics.com/v1/measurements);

• :floor-times? - if true, all timestamps will be floored to :step;

Atlas

Sends metrics to Netflix Atlas server.

Add POM dependency: [io.micrometer/micrometer-registry-atlas "1.5.1"]

(configure
  {:type :atlas,
   :auto-start? true,
   :url "http://atlas.svc.kubernetes.local:7101/api/v1/publish"})

It is a push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout and :batch-size.

Custom settings:

• :url - URL to Atlas server metrics publish endpoint (required, default: http://localhost:7101/api/v1/publish);

• :meter-ttl - time duration after which inactive meters stop reporting (default: 15m);

• :auto-start? - if true, sender threads will start automatically (default: true);

• :lwc-step - reporting frequency for streaming to atlas LWC (must be less or equal than :step, default: 5s);

• :lwc-enabled? - if true, Atlas LWC streaming is enabled (default: false); for more details see Atlas documentation;

• :lwc-ignore-publish-step? - if true if expressions with the same step size as Atlas publishing should be ignored for streaming; for more details see Atlas documentation;

• :config-refresh-frequency - config refresh frequency for LWC service (default: 10s);

• :config-ttl - config validity duration (default: 150s);

• :config-url - LWC config information endpoint URL (default: http://localhost:7101/lwc/api/v1/expressions/local-dev);

• :eval-url - LWC service URL to evaluate data for subscription (default: http://localhost:7101/lwc/api/v1/evaluate);

• :common-tags - map of common tags that will be attached to all metrics;

• :propagate-warnings? - if true, will treat warnings as errors (default: false);

• :max-atlas-metrics - if set, will limit number of distinct metrics to be sent;

• :gauge-polling-frequency - how often registered gauges should be polled (default: 10s);

Azure

Sends data to Azure Monitor service in Azure cloud. Uses client library provided by Microsoft and its configuration properties, so here we set only instrumentation key for monitored application:

Add POM dependency: [io.micrometer/micrometer-registry-azure "1.5.1"]

(configure
  {:type :azure
   :instrumentation-key "25678f44-b958-466e-9ccc-5a622484620b"})

It is push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :instrumentation-key - instrumentation key as configured in azure monitor (required);

Cloudwatch

Sends data to Amazon Cloudwatch service. As part of configuration it needs CloudWatch client created using CloudWatchAsyncClient.create() method.

Add POM dependency: [io.micrometer/micrometer-registry-cloudwatch2 "1.5.1"]

(configure
  {:type :cloudwatch})

It is push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :namespace - cloud watch namespace;

Datadog

Sends data to Datadog service.

Add POM dependency: [io.micrometer/micrometer-registry-datadog "1.5.1"]

(configure 
  {:type :datadog
   :api-key "f7cf4d9d-0755-4e5b-bc3b-fa0b54f14ea6"
   :application-key "my-app.svc.mycompany.com"})

It is push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :api-key - datadog API key (required);

• :application-key - datadog application key;

• :host-tag - tag mapped to host when shipping metrics to datadog;

• :url - URI to ship metrics to (default: https://api.datadoghq.com);

• :descriptions? - when true, meter descriptions will be sent to datadog;

Dynatrace

Sends data to Dynatrace system.

Add POM dependency: [io.micrometer/micrometer-registry-dynatrace "1.5.1"]

(configure
  {:type :dynatrace
   :url "https://url-to-dynatrace-collector"
   :api-token "abcdefjhij1234567890"
   :device-id "..."})

It is push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :api-token - dynatrace API token;

• :url - URL to dynatrace collector;

• :device-id - dynatrace device ID;

• :technology-type - technology type label, eg. java or clojure;

• :group - application group name;

Elastic

Sends data to Elastic Search server.

Add POM dependency: [io.micrometer/micrometer-registry-elastic "1.5.1"]

(configure
  {:type :elastic
   :url "https://elastic.svc.cloud.mycomapny.com:9200"
   :index "metrics"
   :username "micrometer"
   :password "seCR3t"})

It is push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :url - URL to Elastic server storing data (mandatory);

• :index - name of index used to write metrics to (default: metrics);

• :index-date-format - used when rolling indexes, (default: yyyy-MM);

• :index-date-separator - string separating index name from date suffix (default: -);

• :username, :password - credentials to use when authentication is enabled in elastic server;

• :timestamp-field-name - timestamp field name (default: @timestamp);

• :auto-create-index? - if true and index does not exist, it will be created (default: true);

• :pipeline - ingest pipeline name (default: pipeline);

• :document-type - document type (default: doc); note that this feature is deprecated in newer versions of Elastic;

Ganglia

Sends data to Ganglia over UDP.

Add POM dependency: [io.micrometer/micrometer-registry-ganglia "1.5.1"]

(configure
  {:type :ganglia
   :host "224.2.3.4"
   :port 8649})

It is push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :duration-units - time unit for duration values: :nanoseconds, :microseconds, :milliseconds or :seconds, (default value: :milliseconds);

• :addressing-mode - UDP sender addressing mode: :unicast or :multicast (default: :multicast);

• :host, :port - host address (can be multicast address) and port name;

• :ttl - TTL for UDP packets; this is important for multicast communication, eg. when set to 1, data will not be transmitted outside of LAN machine is connected to;

Graphite

Sends data to Graphite server.

Add POM dependency: [io.micrometer/micrometer-registry-graphite "1.5.1"]

(configure
  {:type :graphite
   :host "graphite.intranet.mycompany.com"})

It is push registry and accepts common settings: :step, :enabled?;

Custom settings:

• :host, :port - host name and port number (required);

• :graphite-tags-enabled? - when true, graphite tags will be used rather than hierarchical naming convention (default: true);

• :tags-as-prefix - list of tags used to construct metric prefix when hierarchical convention is used;

• :rate-units - time units used when submitting rate metrics: :nanoseconds, :microseconds, :milliseconds or :seconds (default: :milliseconds);

• :duration-units - time units used when submitting duration metrics: :nanoseconds, :microseconds, :milliseconds or :seconds (default: :milliseconds);

• :protocl - communication protocol: :plaintext, :udp or :pickled (default: :pickled);

Humio

Sends metrics to Humio cloud service.

Add POM dependency: [io.micrometer/micrometer-registry-humio "1.5.1"]

(configure
  {:type :humio
   :api-token "44f575d5-b591-47c7-a01e-71d6771280b6"})

It is push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :api-token - API token for authentication in Humio service;

• :tags - tags which define which datasource to store metrics in (see Humio documentation);

• :url - URL to Humio cloud API service (default: https://cloud.humio.com)

Influx

Sends metrics to InfluxDB time series database.

Add POM dependency: [io.micrometer/micrometer-registry-influx "1.5.1"]

(configure
  {:type :influx
   :url "http://influx.svc.kubernetes.local:8086"
   :username "metrics"
   :password "s3CR3t"})

It is push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :url - URL to influxdb server (required, default: http://localhost:8086);

• :db - database name (default: metrics);

• :auto-create-db? - if set to true, will create database if non-existent (default: true);

• :username, :password - credentials to be used to authenticate to influxdb server;

• :consistency - write consistency for submitted data if influxdb is clustered: :all, :one, :any, :quorum (default: :one);

• :retention-policy - influx retention policy name for stored data (if not set, DEFAULT will be used);

• :retention-duration - retention duration for stored data (eg. 2h or 52w);

• :retention-replication-factor - how many copies of data will be stored (must be 1 for single node instances);

• :retention-shard-duration - time range covered by shard group (eg. 2h, 52w);

• :compressed? - if true, transmitted data will be GZIP compressed (default: true);

JMX

Presents data via JMX in local MBean server.

Add POM dependency: [io.micrometer/micrometer-registry-jmx "1.5.1"]

(configure
  {:type :jmx
   :domain "micrometer"})

Custom settings:

• :domain - JMX domain used in object names of published metrics;

• :step - reporting frequency;

Kairos

Sends metrics to Kairos monitoring system.

Add POM dependency: [io.micrometer/micrometer-registry-jmx "1.5.1"]

(configure
  {:type :kairos
   :url "http://kairos.svc.kubernetes.local:8080/api/v1/datapoints"
   :username "metrics"
   :password "seCR3t"})

It is push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :url - URI to data submission endpoint (default: http://localhost:8080/api/v1/datapoints);

• :username, :password - credentials to authenticate in kairos system (required);

New Relic

Sends metrics to New Relic service.

Add POM dependency: [io.micrometer/micrometer-registry-newrelic "1.5.1"]

(configure
  {:type :newrelic
   :account-id "123456"
   :api-key "9a4186e2-7fd2-48ec-a9a3-c2b29b5fff27"}) 

It is push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :url - URL to insights collector (default: https://insights-collector.newrelic.com);

• :api-key - insigts collector API key (required);

• :account-id - New Relic account ID (required);

• :meter-name-event-type-enabled? - if true, will use meter names as New Relic eventType values (default: false);

• :event-type - default eventType (used if meter names are not used);

• :client-provider-type - either :insights_api or :insights_agent;

OpenTSDB

Sends metrics to OpenTSDB server.

Add POM dependency: [io.micrometer/micrometer-registry-opentsdb "1.5.1"]

(configure
  {:url "http://opentsdb.intranet.mycompany.com:4242/api/put"
   :username "metrics"
   :password "s3CR3t"})

It is push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :url - URI to data submission endpoint (default: http://localhost:4242/api/put);

• :username, :password - credentials to authenticate in opentsdb server;

Prometheus

Add POM dependency: [io.micrometer/micrometer-registry-prometheus "1.5.1"]

Keeps metrics data in memory and provides method to generate Prometheus scrape data. Note that this implementation does not serve HTTP endpoint for Prometheus scrapper, it provides scrape multimethod implementation instead that can be used by application developer who is responsible for publishing this data via http(s).

(configure
  {:type :prometheus})

In order to generate data for prometheus, use scrape multimethod:

; implement this as URI in web handler instead of using println
(println (scrape *registry*))

Custom settings:

• :prefix - string that will be prepended to names of all metrics sent (default: metrics);

• :step - window size when computing timed or distribution statistics (default: 1m);

• :descriptions? - if true, description hints will be generated;

• :histogram-flavor - histogram type backing DistributionSummary and Timer meters: :Prometheus or :VictoriaMetrics, note that these keyword are case sensitive (default: Prometheus);

SignalFX

Sends metrics to SignalFX service.

Add POM dependency: [io.micrometer/micrometer-registry-signalfx "1.5.1"]

(configure
  {:type :signalfx
   :access-token "3d0146bb-56e5-46c0-8768-64508e2914fb"
   :source "myapp"})

It is push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :access-token - access token for SignalFX service (required);

• :url - URL to SignalFX ingestion service (default: https://ingest.signalfx.com);

• :source - application identifier (required, should be unique for each application);

Stackdriver

Sends metrics to Google Stackdriver service.

Add POM dependency: [io.micrometer/micrometer-registry-stackdriver "1.5.1"]

(configure
  {:type :stackdriver
   :project-id "myapp"
   :credentials "/tmp/credentials.dat"})

It is push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :credentials - path to credentials file;

• :project-id - project ID (as reported to stackdriver);

• :resource-labels - map of resource labels; see stackdriver documentation;

• :resource-type - resource type; see stackdriver documentation;

Statsd

Sends metrics to statsd daemon

Add POM dependency: [io.micrometer/micrometer-registry-statsd "1.5.1"]

(configure
 {:type :statsd
   :host "statsd.svc.kubernetes.local"
   :port 8125})

Custom settings:

• :host - statd server address (required);

• :port - statsd service port number (default: 8125);

• :protocol - either :udp or :tcp (default: :udp);

• :enabled? - enables or disables data submission;

• :flavor - one of :etsy, :datadog, :telegraf, :sysdig (default: :datadog);

• :max-packet-length - limits packet length for UDP communication; should be less than MTU between monitored application and statsd server - 1431 for fast ethernet, 8932 for gigabit ethernet, 512 for public internet (default: 1400);

• :polling-frequency - determines how often gauges will be polled (default: 10s);

• :queue-size - send queue size (default: no limits);

• :step - step size to use in computing windowed statistics (default: 60s);

• :publish-unchanged-meters? - if true, unchanged meters will be sent each time (default: true);

• :buffered? - if true, measurements will be buffered and sent when max packet length is reached or until polling frequency is reached (default: true);

Wavefront

Sends metrics to wavefront service.

Add POM dependency: [io.micrometer/micrometer-registry-wavefront "1.5.1"]

(configure
  {:type :wavefront,
   :api-token "e6cb63cd-5b14-4bd4-9486-8ff6cfddd830",
   :source "myapp"})

It is push registry and accepts common settings: :step, :enabled?, :num-threads, :connect-timeout, :read-timeout and :batch-size.

Custom settings:

• :api-token - wavefront API token (required);

• :url - URL to wavefront service (default: https://longboard.wavefront.com);

• :source - application identifier (required);

• :report-minute-distribution? - report histogram distributions into minute intervals;

• :report-hour-distribution? - report histogram distributions into hour intervals;

• :report-day-distributeion? - report histogram distributions into day intervals;