This is the semi-official 'all-in-one' Logstash cookbook.
All of the requirements are explicitly defined in the recipes. Every effort has been made to utilize official Opscode cookbooks.
However if you wish to use an external ElasticSearch cluster, you will need to install that yourself and change the relevant attributes for discovery. The same applies to integration with Graphite.
This cookbook has been tested together with the following cookbooks
- Heavywater Graphite Cookbook - This is the one I use
- Karmi's ElasticSearch Cookbook
node[:logstash][:basedir]- the base directory for all the Logstash componentsnode[:logstash][:user]- the owner for all Logstash componentsnode[:logstash][:group]- the group for all Logstash componentsnode[:logstash][:graphite_role]- the Chef role to search for discovering your preexisting Graphite servernode[:logstash][:elasticsearch_role]- the Chef role to search for discovering your preexisting ElasticSearch cluster.node[:logstash][:elasticsearch_cluster]- the cluster name assigned to your preexisting ElasticSearch cluster. Only applies to external ES clusters.node['logstash']['elasticsearch_ip']- the IP address that will be used for your elasticsearch server in case you are using Chef-solonode['logstash']['graphite_ip']- the IP address that will be used for your graphite server in case you are using Chef-solo
node[:logstash][:agent][:install_method]- The method to install logstash - eitherjarorsource, defaults tojarnode[:logstash][:agent][:version]- The version of Logstash to install. Only applies tojarinstall method.node[:logstash][:agent][:source_url]- The URL of the Logstash jar to download. Only applies tojarinstall method.node[:logstash][:agent][:checksum]- The checksum of the jar file. Only applies tojarinstall method.node[:logstash][:agent][:base_config]- The name of the template to use forlogstash.confas a base config.node[:logstash][:agent][:base_config_cookbook]- Where to find the base_config template.node[:logstash][:agent][:xms]- The minimum memory to assign the JVM.node[:logstash][:agent][:xmx]- The maximum memory to assign the JVM.node[:logstash][:agent][:debug]- Run logstash with-voption?node[:logstash][:agent][:server_role]- The role of the node behaving as a Logstashserver/indexer
node[:logstash][:server][:install_method]- The method to install logstash - eitherjarorsourcenode[:logstash][:server][:version]- The version of Logstash to install. Only applies tojarinstall method.node[:logstash][:server][:source_url]- The URL of the Logstash jar to download. Only applies tojarinstall method.node[:logstash][:server][:checksum]- The checksum of the jar file. Only applies tojarinstall method.node[:logstash][:server][:base_config]- The name of the template to use forlogstash.confas a base config.node[:logstash][:server][:base_config_cookbook]- Where to find the base_config template.node[:logstash][:server][:xms]- The minimum memory to assign the JVM.node[:logstash][:server][:xmx]- The maximum memory to assign the JVM.node[:logstash][:server][:debug]- Run logstash with-voption?node[:logstash][:server][:enable_embedded_es]- Should Logstash run with the embedded ElasticSearch server or not?
node[:logstash][:kibana][:repo]- The git repo to install Kibana from.node[:logstash][:kibana][:sha]- The sha/branch of the repo you wish to clone.node[:logstash][:kibana][:apache_template]- The name of the template file to use for the Apache site filenode[:logstash][:kibana][:config]- The name of the template to use for the Kibanaconfig.phpfilenode[:logstash][:kibana][:server_name]- The value to use for the ApacheServerNamevariable to use for the Kibana Apache virtual host.
node[:logstash][:source][:repo]- The git repo to use for the source code of Logstashnode[:logstash][:source][:sha]- The sha/branch of the repo you wish to clone.node[:logstash][:source][:java_home]- yourJAVA_HOMElocation. Needed explicity forantwhen building JRuby
A proper readme is forthcoming but in the interim....
There are 3 recipes you need to concern yourself with:
- server - This would be your indexer node
- agent - This would be a local host's agent for collection
- kibana - This is the web interface
Every attempt (and I mean this) was made to ensure that the following objectives were met:
- Any agent install can talk to a server install
- Kibana web interface can talk to the server install
- Each component works OOB and with each other
- Utilize official opscode cookbooks where possible
This setup makes HEAVY use of roles. Additionally, ALL paths have been made into attributes. Everything I could think of that would need to be customized has been made an attribute.
By default, the recipes look for the following roles (defined as attributes so they can be overridden):
graphite_server-node[:logstash][:graphite_role]elasticsearch_server-node[:logstash][:elasticsearch_role]logstash_server-node[:logstash][:kibana][:elasticsearch_role]andnode[:logstash][:agent[:server_role]
The reason for giving kibana its own role assignment is to allow you to point to existing ES clusters/logstash installs.
The reason for giving agent its own role assignment is to allow the server and agent recipes to work together.
Yes, if you have a graphite installation with a role of graphite_server, logstash will send stats of events received to logstash.events.
The template to use for configuration is made an attribute as well. This allows you to define your OWN logstash configuration file without mucking with the default templates.
The server will, by default, enable the embedded ES server. This can be overriden as well.
See the server and agent attributes for more details.
Both agent and server support an attribute for how to install. By default this is set to jar to use the 1.1.1preview as it is required to use elasticsearch 0.19.4. The current release is defined in attributes if you choose to go the source route.
Here are some basic steps
- Create a role called
logstash_serverand assign it the following recipes:logstash::serverandlogstash::kibana - Assign the role to a new server
- Assign the
logstash::agentrecipe to another server
If there is a system found with the logstash_server role, the agent will automatically configure itself to send logs to it over tcp port 5959. This is, not coincidently, the port used by the chef logstash handler.
If there is NOT a system with the logstash_server role, the agent will use a null output. The default input is to read files from /var/log/*.log excluding and gzipped files.
If you point your browser to the logstash_server system's ip address, you should get the kibana web interface.
Do something to generate a new line in any of the files in the agent's watch path (I like to SSH to the host), and the events will start showing up in kibana. You might have to issue a fresh empty search.
The pyshipper recipe will work as well but it is NOT wired up to anything yet.
The current templates for the agent and server are written so that you can provide ruby hashes in your roles that map to inputs, filters, and outputs. Here is a role for logstash_server
name "logstash_server"
description "Attributes and run_lists specific to FAO's logstash instance"
default_attributes(
:logstash => {
:server => {
:enable_embedded_es => false,
:inputs => [
:amqp => {
:type => "all",
:host => "127.0.0.1",
:exchange => "rawlogs",
:name => "rawlogs_consumer"
}
],
:filters => [
:grok => {
:type => "haproxy",
:pattern => "%{HAPROXYHTTP}",
:patterns_dir => '/opt/logstash/server/etc/patterns/'
}
],
:outputs => [
:file => {
:type => 'haproxy',
:path => '/opt/logstash/server/haproxy_logs/%{request_header_host}.log',
:message_format => '%{client_ip} - - [%{accept_date}] "%{http_request}" %{http_status_code} ....'
}
]
}
}
)
run_list(
"role[elasticsearch_server]",
"recipe[logstash::server]",
"recipe[php::module_curl]",
"recipe[logstash::kibana]"
)
It will produce the following logstash.conf file
input {
amqp {
name => 'rawlogs_consumer'
exchange => 'rawlogs'
type => 'all'
host => '127.0.0.1'
}
}
filter {
grok {
pattern => '%{HAPROXYHTTP}'
type => 'haproxy'
patterns_dir => '/opt/logstash/server/etc/patterns/'
}
}
output {
stdout { debug => true debug_format => "json" }
elasticsearch { host => "169.1.1.1" }
file {
type => 'haproxy'
message_format => '%{client_ip} - - [%{accept_date}] "%{http_request}" %{http_status_code} %{bytes_read} ....'
path => '/opt/logstash/server/haproxy_logs/%{request_header_host}.log'
}
}
- Currently only tested on Ubuntu Natty, Precise, and RHEL 6.2.
Author:: John E. Vincent Author:: Bryan W. Berry ([email protected]) Author:: Richard Clamp (@richardc) Author:: Juanje Ojeda (@juanje) Copyright:: 2012, John E. Vincent Copyright:: 2012, Bryan W. Berry Copyright:: 2012, Richard Clamp Copyright:: 2012, Juanje Ojeda
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.