Sponsored by Amplifr

Yabeda::Puma::Plugin

Built-in metrics for Puma web server monitoring out of the box! Part of the yabeda suite.

Metrics

Works as the Puma plugin and provides following metrics:

  • puma_workers - the number of running puma workers
  • puma_booted_workers - the number of booted puma workers
  • puma_old_workers - the number of old puma worker

Segmented by the worker (index of the worker):

  • puma_pool_capacity - the capacity of each worker: the number of requests that the server is capable of taking right now. More details are here.
  • puma_running - the number of running threads (spawned threads) for any puma worker
  • puma_max_threads - preconfigured maximum number of worker threads
  • puma_backlog - the number of backlog threads, the number of connections in that worker's "todo" set waiting for a worker thread.

Installation

Add this line to your application's Gemfile:

gem 'yabeda-puma-plugin'

And then execute:

$ bundle

Usage

Collecting metrics

Add those 2 lines of code to your config/puma.rb file:

activate_control_app
plugin :yabeda

It will activate default puma control application working over the unix socket, and runs the yabeda puma plugin, for registering and collecting the metrics.

Exposing metrics

Some monitoring system agents (like NewRelic or DataDog) will send metrics automatically in the background. But for some of monitoring systems (like Prometheus) you have to explicitly set up metrics export.

Prometheus

On the same endpoint with your application

For non-Rails applications place following line in your config.ru before running your application:

use Yabeda::Prometheus::Exporter, path: "/metrics"

In Ruby on Rails applications you can add following line in config/routes.rb instead:

mount Yabeda::Prometheus::Exporter => "/metrics"

In both cases your Puma instance metrics (along with your application metrics) will be available at /metrics endpoint.

On different port

Sometimes you don't want to expose metrics publicly for security reasons. For that case prometheus exporter plugin is bundled with this gem.

Don't forget to add either yabeda-prometheus or yabeda-prometheus-mmap gem into your Gemfile!

Add this plugin into your config/puma.rb:

plugin :yabeda_prometheus

By default metrics will be available at http://0.0.0.0:9394/metrics.

Bind host, port, and path can be controlled either by config file option prometheus_exporter_url:

# config/puma.rb
prometheus_exporter_url "tcp://127.0.0.1:9395/shmetrics"

Or by environment variables PROMETHEUS_EXPORTER_URL, PROMETHEUS_EXPORTER_BIND, PROMETHEUS_EXPORTER_PORT, and PROMETHEUS_EXPORTER_PATH (takes precedence over configuration option).

Details

In accordance with the architecture of the puma web server lets look how it works:

Yabeda::Puma get metrics from puma control app over the unix socket.

For the configuration above, we will have the list of metrics (with help of yabeda-prometheus exporter):

GET /metrics

puma_backlog{index="0"} 0
puma_backlog{index="1"} 0
puma_running{index="0"} 5
puma_running{index="1"} 5
puma_pool_capacity{index="0"} 1
puma_pool_capacity{index="1"} 5
puma_max_threads{index="0"} 5
puma_max_threads{index="1"} 5
puma_workers 2
puma_booted_workers 2
puma_old_workers 0

See also the grafana screenshot of monitoring puma pool size and it's capacity when application is overloaded:

Monitor puma metrics with grafana.

Roadmap (TODO or Help wanted)

  • Collect also control-gc puma metrics

Development with Docker

Get local development environment working and tests running is very easy with docker-compose:

docker-compose run app bundle
docker-compose run app bundle exec rspec

Contributing

Bug reports and pull requests are welcome on GitHub at https://github.com/yabeda-rb/yabeda-puma-plugin.

License

The gem is available as open source under the terms of the MIT License.