Overview

Chef cookbook for the monit monitoring and management tool. This cookbook has been updated to configure services defined in a data bag.

This cookbook was forked from StudyBlue/monit, version 0.7.

Attributes

• default[:monit][:services] - List of items from the monit data bag that describe services monitored by monit. For more info, see Usage below.
• default[:monit][:logfile] - The file to which monit logs are written
• default[:monit][:notify][:enable] - enable email notifications
• default[:monit][:notify][:email] - the email to which notifications are sent.
• default[:monit][:httpd][:enable] - enable the http server
• default[:monit][:httpd][:port] - port on which the http server listens
• default[:monit][:httpd][:address] - the address to which the server binds
• default[:monit][:httpd][:allow] - the addresses from which connections are allowed (default %w{localhost})
• default[:monit][:httpd][:signature] - either "enable" or "disable". Determines whether monit shows a signature
• default[:monit][:httpd][:basic_auth_accounts] - An array of admin:"password" entries that are used for HTTP Basic Auth. Not the quotes around the password; these are required.
• default[:monit][:poll_period] - the time (in seconds) between poll cycles
• default[:monit][:poll_start_delay] - the amount of time (in seconds) before monit starts polling services.
• default[:monit][:mail][:server] - host on which the smtp mail server runs
• default[:monit][:mail][:port] - The port on which the SMTP server runs (default is 25)
• default[:monit][:mail][:username] - Username for an SMTP account
• default[:monit][:mail][:password] - Password for an SMTP account
• default[:monit][:mail][:security] - Send secure email. Must be one of: SSLV2, SSLV3 or TLSV1
• default[:monit][:mail][:checksum] - SMTP Server certificat checksum (optional)
• default[:monit][:mail][:timeout] - The timeout (in seconds) for sending mail
• default[:monit][:mail][:use_node_fqdn] - Whether or not to use the node's FQDN in as the hostname (default is false)
• default[:monit][:mail][:format][:subject] - default subject for emails from monit
• default[:monit][:mail][:format][:from] - email address from which messages are sent
• default[:monit][:mail][:format][:message] - template for messages (see attributes/default.rb for details)
• default[:monit][:queue][:location] - base directory where events will be stored
• default[:monit][:queue][:slots] - limit the size of the queue; this is the number of messages that will be held

Usage

Right now, this cookbook only supports simple process monitoring. To set up process monitors, create a monit data bag, and create items that look similar to the following:

This would be data_bags/monit/monit_services.json:

{
  "id": "monit_services",
  "postgresql": {
    "process_name": "postgres",
    "pidfile":      "/var/run/postgresql/9.1-main.pid",
    "start_command":"/usr/sbin/service postgresql start",
    "stop_command": "/usr/sbin/service postgresql stop",
    "check_port": "5432",
    "check_protocol": "pgsql",
    "check_times": 2,
    "within_cycles": 3,
    "then_action": "restart"
  },
  "supervisor": {
    "process_name": "supervisor",
    "pidfile":      "/var/run/supervisord.pid",
    "start_command":"/etc/init.d/supervisor start",
    "stop_command":"/etc/init.d/supervisor stop",
    "check_socket": "/var/run/supervisor.sock",
    "check_times": 1,
    "within_cycles": 2,
    "then_action": "restart"
  }
}

This item would set up a monitor for the postgres and supervisor processes. Note that it checks for PostgreSQL using a TCP port while Supervisor uses a Unix socket.

This would result in the following:

• /etc/monit/conf.d/postgresql.conf:

    CHECK PROCESS postgres WITH PIDFILE /var/run/postgresql/9.1-main.pid
      START PROGRAM "/usr/sbin/service postgresql start"
      STOP PROGRAM "/usr/sbin/service postgresql stop"
      IF FAILED PORT 5432 PROTOCOL pgsql 2 TIMES WITHIN 3 CYCLES THEN restart
  

• /etc/monit/conf.d/supervisor.conf:

    CHECK PROCESS supervisor WITH PIDFILE /var/run/supervisord.pid
      START PROGRAM "/etc/init.d/supervisor start"
      STOP PROGRAM "/etc/init.d/supervisor stop"
      IF FAILED unixsocket /var/run/supervisor.sock 1 TIMES WITHIN 2 CYCLES THEN restart
  

You can then set up a role, and override any of the existing attributes, and specify any data bag items taht will be used to set up service monitoring.

You might have a monit role that looks like:

name "monit"
description "Monit role"
run_list(
  "recipe[monit::default]",
  "recipe[monit::services]"
)

default_attributes(
  "monit" => {
    "services" => ["monit_services", ],
  }
)

History

version 0.8.1

• added additional SMTP configuration
• added additional HTTPD configuration

version 0.8

• added more generic services recipes that pulls configuration data from a data bag
• only supporting Ubuntu/Debian
• removed postfix/ssh recipes

version 0.7

• create /etc/monit/conf.d. Thanks Karel Minarik (https://github.com/karmi)

version 0.6

• Released to github
• Defaults no alert on ACTION event. When you manually stop/start a service, alerting me about what I just did isn't useful.