eye is a process monitoring tool, inspired by bluepill and god. This chef cookbook installs the eye RubyGem and configures it to manage services. It also includes a LWRP.
eye is written in Ruby, so it should work on any system. If you want to use eye's uid
or gid
features, Ruby > 2.0.0 is required.
node['eye']['version']
- Version of eye, default is '0.6.1'node['eye']['conf_dir']
- Location of eye's configuration files, default '/etc/eye'node['eye']['bin']
- Location of eye binary, default "node['languages']['ruby']['bin_dir']/eye"node['eye']['log_dir']
- Location of eye's logfiles, default '/var/log/eye'node['eye']['run_dir']
- Path of state directory, default '/var/run/eye'node['eye']['user']
- Owner of eye process, directories and configuration files, default 'root'node['eye']['group']
- Group of eye process, directories and configuration files, default 'root'node['eye']['http']['install']
- if set to true,eye-http
gem will be installed, default 'false'node['eye']['http']['version']
- Version of eye-http, default is '0.6'
This cookbook contains a LWRPs, eye_service.
eye_service 'my_service' do
action [:enable, :reload, :restart]
end
This loads eye daemon with a needed configuration 'my_service.eye' in a subdirectory of node['eye']['conf_dir']
named by the default eye owner (e.g /etc/eye/root/my_service.rb) and starts the according process. The owner of eye daemon and the process itself is the user defined with node['eye']['user']
An (additional) eye daemon can also be started by a different owner (with it's own logfile, e.g in '/var/log/eye/deploy/eye.log')
eye_service 'my_service' do user_srv true user_srv_uid 'deploy' user_srv_gid 'deploy' user_srv_home '/home/deploy' action [:enable, :reload, :restart] end
In that case the configuration file should be in '/etc/eye/deploy/my_service.eye'
If eye daemon should by controlled by node['eye']['user']
, but the process should be started as different user, use eye's uid
or gid
configuration features.
eye_service
creates an init script, e.g. '/etc/init.d/my_service'. This init script can be prefixed, like
eye_service 'my_service' do
init_script_prefix 'eye-'
...
end
There's also a definition eye_app
which creates the necessary configuration files, log directories and calls eye_service
afterwards.
Example for unicorn:
eye_app "test_unicorn" do
user_srv true
user_srv_uid "deploy"
user_srv_gid "deploy"
user_srv_home '/home/deploy'
enable true
reload false
restart false
template "rails_unicorn.conf.erb"
cookbook "rails_app"
variables :ruby => "#{node['languages']['ruby']['bin_dir']}/ruby",
:environment => 'test',
:working_dir => '/var/www/rails_dir'
end
with according template:
Eye.application "test_unicorn" do
env "RAILS_ENV" => '<%= @environment %>'
env "PATH" => "#{File.dirname("<%= @ruby %>")}:#{ENV['PATH']}"
working_dir "<%= @working_dir %>"
process("unicorn") do
pid_file "tmp/pids/unicorn.pid"
start_command "<%= @ruby %> ./bin/unicorn -Dc ./config/unicorn.rb -E <%= @environment %>"
stdall "log/unicorn.log"
stop_signals [:TERM, 10.seconds]
restart_command "kill -USR2 {PID}"
check :cpu, :every => 30, :below => 80, :times => 3
check :memory, :every => 30, :below => 150.megabytes, :times => [3,5]
start_timeout 30.seconds
restart_grace 30.seconds
monitor_children do
stop_command "kill -QUIT {PID}"
check :cpu, :every => 30, :below => 80, :times => 3
check :memory, :every => 30, :below => 150.megabytes, :times => [3,5]
end
end
end
e.g.
Just include eye
in your node's run_list
:
{
"name":"my_node",
"run_list": [
"recipe[eye]"
]
}
- Fork the repository on Github
- Create a named feature branch (like
add_component_x
) - Write you change
- Write tests for your change (if applicable)
- Run the tests, ensuring they all pass
- Submit a Pull Request using Github
Authors: Holger Amann holger@fehu.org
Licensed under Apache License, Version 2.0 http://www.apache.org/licenses/LICENSE-2.0