Installs and configures Munin for a server and for clients using Chef Server search capabilities.
Requires Chef 0.10.0+ for Chef environments. Supports both chef-solo and chef-client ( with node searches support ).
Chef-solo:
- Since there is no search support for chef-solo, current node object will be used instead.
- See Data Bags section below in order to configure access to Munin web UI.
Chef-client:
- The monitoring server that uses this recipe should have a role named '
munin_monitoring
'. This is configurable with an attribute (see below). The recipes use search, and narrow the results to nodes in the samechef_environment
.
- Debian/Ubuntu
- Red Hat 5.7, 6.1
- ArchLinux
- apache2 - Not required (users can simply be in a data bag and the recipe will search), but useful for setting the openid list (see OpenID Authentication, below).
- users (see Data Bags) - Not required, but recommended to install perl cpan modules for munin plugins
- perl
- nginx
node['munin']['web_server_port']
- port that the munin vhost runs on, default 80node['munin']['sysadmin_email']
- default email address for serveradmin in vhost.node['munin']['server_auth_method']
- the authentication method to use, default is openid. Any other value will use htauth basic with an htpasswd file.node['munin']['multi_environment_monitoring']
- allow multi-environment monitoring. Default is false. Allowed values are 'true', 'false' or a list of names of chef-environments.node['munin']['server_role']
- role of the munin server. Default is munin_monitoring.node['munin']['docroot']
- document root for the server apache vhost. on archlinux, the default is/srv/http/munin
, or/var/www/munin
on other platforms.node['munin']['web_server']
- supports apache or nginx, default is "apache"node['munin']['public_domain']
- override munin domain.node['munin']['graph_strategy']
- choose a graphing strategy. Default is unset, meaning munins default.cgi
does graph-generation on demand by cgi.node['munin']['max_processes']
- Maximum number of simultaneous Munin-update processes. When not set, munin will use as many as necessary. Default is to use as many as necessary.node['munin']['max_graph_jobs']
- Maximum number of parallel processes used by munin-graph when calling rrdgraph, default is "6"node['munin']['max_cgi_graph_jobs']
- Maximum number of parallel munin-cgi-graph or munin-fastcgi-graph jobs, default is "6"
The client recipe installs munin-node package and starts the service. It also searches for a node with the role for the munin server, by default node['munin']['server_role']
. On Archlinux, it builds the list of plugins to enable.
The server recipe will set up the munin server with Apache. It will create a cron job for generating the munin graphs, search for any nodes that have munin attributes (node['munin']
), and use those nodes to connect for the graphs.
Optionally the server can be setup with NGINX.
Create a users
data bag that will contain the users that will be able to log into the Munin webui. Each user can use htauth with a specified password, or an openid. Users that should be able to log in should be in the sysadmin group. Example user data bag item:
{
"id": "munin",
"groups": "sysadmin",
"htpasswd": "hashed_htpassword",
"openid": "http://munin.myopenid.com/"
}
When using server_auth_method
'openid', use the openid in the data bag item. Any other value for this attribute (e.g., "htauth", "htpasswd", etc) will use the htpasswd value as the password in /etc/munin/htpasswd.users
.
- The openid must have the http:// and trailing /*. See OpenID Authentication below for more information.
The htpasswd must be the hashed value. Get this value with htpasswd:
% htpasswd -n -s munin
New password:
Re-type new password:
nagiosadmin:{SHA}oCagzV4lMZyS7jl2Z0WlmLxEkt4=
For example use the {SHA}oCagzV4lMZyS7jl2Z0WlmLxEkt4=
value in the data bag.
Create a role named munin_monitoring
that includes the munin::server recipe in the run list. Adjust the docroot to suit your environment.
name 'munin_monitoring'
description 'The monitoring server'
run_list(
'recipe[munin::server]'
)
Apply this role to a node and it will be the munin server. Optionally create a DNS entry for it as munin, that will be used in the Apache vhost.
Use Chef 0.10's environments. For example, create a "production" environment Ruby DSL file and upload it to the Chef Server
name 'production'
description 'Nodes in production'
For chef-client mode clients will automatically search for the server based on the value of the node['munin']['server_role']
attribute in the same environment. If you don't use munin_monitoring
as the role name, change it in a role that is applied to any nodes that get the munin::client
recipe.
By default munin-server will only search the nodes own chef-environment for clients to monitor. Monitoring all available clients in all environment can be done by setting the node['munin']['multi_environment_monitoring'] = 'true'
.
When node['munin']['multi_environment_monitoring']
is set to a list of environment-names, only these environments are searched for munin-clients to monitor. So if you have the environments 'prod', 'staging' and 'dev' and want only nodes in 'prod' and 'staging' to be monitoried by munin-server, set node['munin']['multi_environment_monitoring'] = ['prod', 'staging']
.
The recipe apache2::mod_auth_openid
is updated to a version of the module that apparently does not support the AuthOpenIDUserProgram
directive anymore. The virtual host file has been updated to use the Apache HTTPD require user
directive, with a concatenated list from node['apache']['allowed_openids']
. This value must be an array of OpenIDs. Use of the users::sysadmins
recipe will set this up based on data bag search results.
This section describes how to add custom munin plugins.
The munin cookbook now has a definition that can be used to enable a new plugin for data gathering on a client. If an existing munin plugin is desired, call the definition
munin_plugin 'nfs_client'
By default the plugin file name is the name parameter here. Specify the plugin parameter to use something else.
munin_plugin 'nfs_client'
plugin 'nfs_client_'
end
This creates a symlink from the plugins distribution directory, /usr/share/munin/plugins
to the enabled plugins directory, /etc/munin/plugins
, and once the server poller picks it up will have new graph data for that plugin. See the plugins distribution directory for available plugins or add your own.
If a custom plugin is required, add the plugin file to the munin cookbook in site-cookbooks/munin/files/default/plugins
. Call the definition specifying that the plugin file should be downloaded from the cookbook.
munin_plugin 'nfs_client_custom'
create_file true
end
By default in both cases, the plugin is enabled. If a plugin should be disabled, use the enable
parameter
munin_plugin 'nfs_client_custom'
enable false
end
Some plugins may require other configuration. For example, to use the memcache plugins, you'll need the Cache::Memcache
cpan module installed, and use the munin_plugin
definition. The perl cookbook from opscode includes a definition to handle this easily.
cpan_module 'Cache::Memcached'
Then for example in your memcache recipe
%w(
memcached_bytes_
memcached_connections_
memcached_hits_
memcached_items_
memcached_requests_
memcached_responsetime_
memcached_traffic_
).each do |plugin_name|
munin_plugin plugin_name do
plugin "#{plugin_name}#{node['ipaddress'].gsub('.', '_')}_#{node[:memcached][:port]}"
create_file true
end
end
- Author:: Nathan Haneysmith [email protected]
- Author:: Joshua Timberman [email protected]
Copyright 2009-2013, Opscode, Inc
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.