Prometheus.io client library powered by prometheus.erl

Prometheus.ex is a thin, mostly macro-based wrapper around prometheus.erl. While it's pretty straightforward to use prometheus.erl from Elixir, you might prefer prometheus.ex because it gives you:

• native Elixir syntax;
• native Elixir exceptions;
• configuration helpers that are really handy if you plan to write your custom instrumenter.

defmodule ExampleInstrumenter do
  use Prometheus ## require common Prometheus modules, also alias metrics.

  def setup do
    Histogram.new([name: :http_request_duration_milliseconds,
                   labels: [:method],
                   buckets: [100, 300, 500, 750, 1000],
                   help: "Http Request execution time"])
  end

  def instrument(%{time: time, method: method}) do
    Histogram.observe([name: :http_request_duration_milliseconds,
                       labels: [method]],
                      time)
  end
end

Integrations

• Ecto Instrumenter;
• Elixir plugs Instrumenters and Exporter;
• Fuse plugin
• OS process info Collector (linux-only);
• Phoenix Instrumenter;
• RabbitMQ Exporter.

Erlang VM Collectors

• Memory Collector;
• Statistics Collector;
• System Information Collector.

API

API can be grouped like this:

Standard Metrics & Registry

• Prometheus.Metric.Counter - counter metric, to track counts of events or running totals;
• Prometheus.Metric.Gauge - gauge metric, to report instantaneous values;
• Prometheus.Metric.Histogram - histogram metric, to track distributions of events;
• Prometheus.Metric.Summary - summary metric, to track the size of events;
• Prometheus.Registry - working with Prometheus registries.

All metrics created via new/1 or declare/1 macros. The difference is that new/1 actually wants metric to be new and raises Prometheus.MFAlreadyExistsError if it isn't.

Both new/1 and declare/1 accept options as Keyword. Common options are:

• name - metric name, can be an atom or a string (required);
• help - metric help, string (required);
• labels - metric labels, label can be an atom or a string (default is []);
• registry - Prometheus registry for the metric, can be any term. (default is :default)

Histogram also accepts buckets option. Please refer to respective modules docs for the more information.

General Helpers

• Prometheus.Buckets - linear or exponential bucket generators;
• Prometheus.Contrib.HTTP - helpers for HTTP instrumenters.

Integration Helpers

• Prometheus.Config - provides standard configuration mechanism for custom instrumenters/exporters.

Exposition Formats

• Prometheus.Format.Text - renders metrics for a given registry (default is :default) in text format;
• Prometheus.Format.Protobuf - renders metrics for a given registry (default is :default) in protobuf v2 format.

Advanced

You will need this modules only if you're writing custom collector for app/lib that can't be instrumented directly.

• Prometheus.Collector - exports macros for managing/creating collectors;
• Prometheus.Model - provides API for working with underlying Prometheus models. You'll use that if you want to create custom collector.