stalin

Given Ruby's proclivity for heap fragmentation, Web application worker processes tend to exhaust all available memory unless the server is restarted periodically. When all of the workers restart at once, downtime or bad request throughput may result.

Stalin is a gem that gracefully kills your workers before they cause swapping, resulting in better availability for your application. Its design goals are modularity and compatibility with a range of platforms and app servers.

Metrics include:

• Resident Set Size (RSS)

Data sources include:

• Linux ProcFS
• Generic ps

Supported app servers include:

• Rainbows (via Rack middleware + SIGQUIT)
• Unicorn (via Rack middleware + SIGQUIT)
• Puma (via Rack middleware + SIGTERM/SIGUSR2 depending on execution model)

Servers known NOT to work:

• Thin (no supervisor process; need an adapter that execs or something)

As you can see, we are far short of our goal to support many servers! More to come as needed; let me know what you need!

Installation

Just include stalin in your Gemfile.

gem 'stalin'

Usage

Decide on which application server you will use. Add some lines to your config.ru to install a suitable Stalin middleware.

# Gem that kills app processes when their heap becomes too fragmented.
require 'stalin'

# Max memory size (RSS) per worker
mb = 1024**2

# Or Stalin::Adapter::Puma
use Stalin::Adapter::Unicorn, (192*mb), (256*mb)

If you instantiate Stalin::Adapter::Rack directly, you have two choices:

• Pass three parameters (app, graceful-shutdown signal and abrupt-shutdown signal) to decide on signalling behavior yourself
• Pass one parameter (app) to let Stalin decide which adapter to use based on which server is resident in memory

Instantiating the Rack middleware with one parameter is deprecated; we'd much rather you be explicit about your application server than rely on our heuristic!

Tuning

TODO

Special Thanks

• @kzk for the [unicorn-worker-killer] gem which this is derived from