crometheus

This a github fork of ezrast's Crometheus with patches that allow it to work with the latest Crystal version (1.4 for now)

Crometheus is a Prometheus client library for instrumenting programs written in the Crystal programming language. For the most part, Crometheus assumes a basic familiarity with Prometheus. To that end, readers may wish to skim the official documentation on Prometheus' data model, metric types, and text exposition format.

Crometheus is in early development and comes with no guarantees. This project is not affiliated with or endorsed by Prometheus.

For latest updates, see CHANGELOG.md.

Installation

Add this to your application's shard.yml:

dependencies:
  crometheus:
    github: darwinnn/crometheus
    branch: master

Usage

require "crometheus"

# Create an unlabeled summary.
summary = Crometheus::Summary.new(
  :my_first_summary,
  "A sample summary metric")

# Observe some values.
summary.observe 100
summary.observe 200.0

# Create a gauge with labels "foo" and "bar".
gauge = Crometheus::Gauge[:foo, :bar].new(
  :my_first_gauge,
  "A sample gauge metric, with labels")

# In some cases the above syntax will cause type inference to fail;
# work around it with the `Crometheus.alias` macro like this.
Crometheus.alias WidgetCounter = Crometheus::Counter[:kind]
widget_counter = WidgetCounter.new(
  :widgets_made,
  "Number of widgets produced")

# Set some values.
gauge[foo: "Hello", bar: "Anthony"].set 3.14159
gauge[foo: "Goodbye", bar: "Clarice"].set -8e12
widget_counter[kind: "sprocket"].inc
7.times{ widget_counter[kind: "pinion"].inc }

# Access the default registry and start up the server.
Crometheus.default_registry.run_server

Then visit http://localhost:5000 to see your metrics (you may see some default process metrics as well):

# HELP my_first_gauge A sample gauge metric, with labels
# TYPE my_first_gauge gauge
my_first_gauge{foo="Hello", bar="Anthony"} 3.14159
my_first_gauge{foo="Goodbye", bar="Clarice"} -8000000000000.0
# HELP my_first_summary A sample summary metric
# TYPE my_first_summary summary
my_first_summary_count 2.0
my_first_summary_sum 300.0
# HELP widgets_made Number of widgets produced
# TYPE widgets_made counter
widgets_made{kind="sprocket"} 1.0
widgets_made{kind="pinion"} 7.0

The above is all the setup you need for straightforward use cases; all that's left is creating real metrics and instrumenting your code. See the reference documentation for the Gauge, Counter, Histogram, and Summary classes to learn more about the available metric types. The bracket notation Gauge[:foo, :bar] in the example above is a bit of macro magic that creates a LabeledMetric with Gauge as a type parameter. See the examples directory for more samples.

For server configuration see the Registry class documentation. If you want to use multiple registries, e.g. to expose two different sets of metrics on different ports, you'll need to instantiate a second Registry object (other than the default) and pass it as a third argument to your metric constructors (after the name and docstring).

If you want to define a custom metric type, see the documentation for the Metric class, and inherit from that.

Alternately, you can just dive into the API Documentation right from the top.

Contributing

1. Fork it
2. Create your feature branch (git checkout -b my-new-feature)
3. Commit your changes (git commit -am 'Add some feature')
4. Push to the branch (git push origin my-new-feature)
5. Create a new Merge Request

Author

• Ezra Stevens - original author
• Darwin - github fork and patches for Crystal 0.34.0