puppet_collectd

Table of Contents

1. Overview
2. Setup - The basics of getting started with collectd
   • What collectd affects
3. Usage - Configuration options and additional functionality
4. Supported Platforms

Overview

This Puppet module installs the collectd from SignalFx. It also configures the installed collectd to send metrics to SignalFx.

With this module you can also configure collectd plugins (e.g. collectd-rabbitmq, collectd-elasticsearch, collectd-redis, etc.) to send metrics to SignalFx.

Setup

puppet module install signalfx/collectd

What collectd affects

This module installs and configures collectd on your system to send various metrics to SignalFx. Be careful if you already have a working collectd as it will replace your existing collectd configuration.

Usage

class { 'collectd' :
    signalfx_api_token  => 'YOUR_SIGNALFX_API_TOKEN'
}

Other valid parameters are (check the params.pp file for default values):

Plugin Interval

You can set the plugin-specific interval for collecting metrics. This overrides the global interval setting. The following example sets the interval for the cpu and disk plugins to 1 second.

class { 'collectd' :
    signalfx_api_token  => 'YOUR_SIGNALFX_API_TOKEN',
    loadplugins => {
        'cpu' => {
          'Interval' => 1
        },
        'disk' => {
          'Interval' => 1
        }
    }
}

Metric Filtering

Plugin Metric Filtering

You can control which plugin-specific metrics are sent to SignalFx using collectd filtering.

The basic usage is to set filter_metrics to true and then add rules to filter_metric_rules based on Type and/or TypeInstance. A rule's name can be anything, Type must be the complete name of type, and TypeInstance can be a regular expression.

filter_metrics => true,
filter_metric_rules => {
  'some_rule_name' => {
    'Type' => 'some_type',
    'TypeInstance' => '^some_type_instance*'
  }
}

The following example filters the metrics sent from the mysql plugin for all types of "mysql_octets" and also the "threads.running" metric.

class { 'collectd::plugins::mysql' :
  modules => {
    'mydb_plugin_instance' => {
      'Host' => '"localhost"',
      'User' => '"admin"',
      'Password' => '"root"',
      'Database' => '"mydb"',
      'Socket' => '"/var/run/mysqld/mysqld.sock"'
    }
  },
  filter_metrics => true,
  filter_metric_rules => {
  'Mysql-Octets-Whitelist' => {
     'Type' => 'mysql_octets'
   },
   'Mysql-RunningThreads-Whitelist' => {
      'Type' => 'threads',
      'TypeInstance' => 'running'
    }
  }
}

Note: This feature is not available for java plugins which use their own mbean filtering technique.

Default Plugin Metric Filtering

You can also control which default plugin metrics are sent to SignalFx using collectd filtering. See the collectd.conf template for the list of default plugins.

The basic usage is to set filter_default_metrics to true and then add rules to filter_default_metric_rules for one or more plugins based on Type and/or TypeInstance. The plugin and rule names are required. The Type must be the complete name of type, and TypeInstance can be a regular expression.

filter_default_metrics => true,
filter_default_metric_rules => {
  'plugin_name' => {
    'rule_name' => {
      'Type' => 'some_type',
      'TypeInstance' => '^some_type_instance*'
    }
  }
}

The following example applies a filter to the default vmem plugin so that only vmpage_io.swap.in and vmpage_io.swap.out metrics are sent.

class { 'collectd' :
  signalfx_api_token  => 'YOUR_SIGNALFX_API_TOKEN',
  filter_default_metrics => true,
  filter_default_metric_rules => {
    'vmem' => {
      'whitelist_vmpage_io' => {
        'Type' => 'vmpage_io',
        'TypeInstance' => '^swap*'
      }
    }
  }
}

Note: This feature will override any default filters in the filtering.conf file.

Supported list of plugins

You may specify parameters on a per-plugin basis. Please check the notes under each plugin.

1. Apache
2. Cassandra
3. Df
4. Disk
5. Docker
6. Elasticsearch
7. HAProxy
8. Iostat
9. Kafka
10. Memcached
11. Mesos
12. MongoDB
13. MySQL
14. Nginx
15. Postgresql
16. RabbitMQ
17. Redis
18. Varnish
19. Vmstat
20. Zookeeper

####Class: collectd::plugin::apache

class { 'collectd::plugins::apache':
  modules => {
    'myinstance' => {
        'URL' => '"http://localhost/mod_status?auto"',
    }
  }
}

See collectd-apache for configurable parameters and apache configuration instructions.

####Class: collectd::plugin::cassandra

class { 'collectd::plugins::cassandra' :
  modules => {
    'connection1' => {
      'ServiceURL' => '"service:jmx:rmi:///jndi/rmi://localhost:7199/jmxrmi"',
      'Host' => '"testcassandraserver[hostHasService=cassandra]"',
      'collect_metrics' => [
        'classes',
        'garbage_collector',
        'memory-heap',
        'memory-nonheap',
        'memory_pool',
        'threading',
        'cassandra-client-read-latency',
        'cassandra-client-read-timeouts',
        'cassandra-client-read-unavailables',
        'cassandra-client-rangeslice-latency',
        'cassandra-client-rangeslice-timeouts',
        'cassandra-client-rangeslice-unavailables',
        'cassandra-client-write-latency',
        'cassandra-client-write-timeouts',
        'cassandra-client-write-unavailables',
        'cassandra-storage-load',
        'cassandra-storage-hints',
        'cassandra-storage-hints-in-progress',
        'cassandra-compaction-pending-tasks',
        'cassandra-compaction-total-completed',
      ]
    }
  }
}

See collectd-cassandra for configurable parameters.

####Class: collectd::plugin::df

Note: This plugin is automatically included with a default configuration. You need to disable the use of the default df plugin before adding this configuration (see use_default_df_plugin in collectd parameters).

class { 'collectd::plugins::df':
  modules => {
    'config' => {
      'ReportInodes'  => 'true',
      'FSType'  => '"tmpfs"',
      'FSType'  => '"devtmpfs"',
      'IgnoreSelected'  => 'true',
    }
  }
}

See collectd-df for configurable parameters and df configuration instructions.

####Class: collectd::plugin::disk

Note: This plugin is automatically included with a default configuration. You need to disable the use of the default disk plugin before adding this configuration (see use_default_disk_plugin in collectd parameters).

class { 'collectd::plugins::disk':
  modules => {
    'config' => {
      'Disk'  => '"sda"',
      'IgnoreSelected'  => 'false',
    }
  }
}

See collectd-disk for configurable parameters and disk configuration instructions.

####Class: collectd::plugin::docker

class { 'collectd::plugins::docker':
  modules => {
    'dockerplugin' => {
        'BaseURL'              => '"unix://var/run/docker.sock"',
        'Timeout'              => '3',
        'Verbose'              => false
    }
  },
  filter_metrics => true,
  filter_metric_rules => {
    'CpuUsage' => {
       'Type' => 'cpu.usage'
     },
    'MemoryUsage' => {
       'Type' => 'memory.usage'
    },
    'NetworkUsage' => {
       'Type' => 'network.usage'
    },
    'BlockIO' => {
       'Type' => 'blkio',
       'TypeInstance' => '^io_service_bytes_recursive-.*'
    }
  }
}

See collectd-docker for configurable parameters.

####Class: collectd::plugin::elasticsearch

class { 'collectd::plugins::elasticsearch':
  modules => {
    'elasticsearch_collectd' => {
        'Verbose'              => false,
        'Cluster'              => '"elasticsearch"',
        'Indexes'              => '["_all"]',
        'EnableIndexStats'     => true,
        'EnableClusterHealth'  => true,
        'Interval'             => 10,
        'IndexInterval'        => 300,
        'DetailedMetrics'      => false,
        'ThreadPools'          => '["search","index"]',
        'AdditionalMetrics'    => '[""]',
    }
  }
}

See collectd-elasticsearch for configurable parameters. The sample output file generated would look like 20-elasticsearch.conf. Currently, the plugin only monitors one elasticsearch instance, so you should include only one module in the above class arguments.

####Class: collectd::plugin::kafka

class { 'collectd::plugins::kafka' :
  modules => {
    'connection1' => {
      'ServiceURL' => '"service:jmx:rmi:///jndi/rmi://localhost:7099/jmxrmi"',
      'Host' => '"testkafkaserver[hostHasService=kafka]"',
      'collect_metrics' => [
        'classes',
        'garbage_collector',
        'memory-heap',
        'memory-nonheap',
        'memory_pool',
        'threading',
        'kafka-all-messages',
        'kafka-all-bytes-in',
        'kafka-all-bytes-out',
        'kafka-log-flush',
        'kafka-active-controllers',
        'kafka-underreplicated-partitions',
        'kafka-request-queue',
        'kafka.fetch-consumer.total-time',
        'kafka.fetch-follower.total-time',
        'kafka.produce.total-time',
      ]
    }
  }
}

See collectd-kafka for configurable parameters.

####Class: collectd::plugin::iostat

class { 'collectd::plugins::iostat':
  modules => {
    'collectd_iostat_python' => {
        'Path' => '"/usr/bin/iostat"',
        'Verbose' => false,
        'Include' => '["tps", "kB_read/s", "kB_wrtn/s", "kB_read", "kB_wrtn", "rrqm/s", "wrqm/s", "r/s", "w/s", "rsec/s", "rkB/s", "wsec/s", "wkB/s", "avgrq-sz", "avgqu-sz", "await", "r_await", "w_await", "svctm", "%util"]'
    }
  }
}

See collectd-iostat for configurable parameters.

####Class: collectd::plugin::memcached

class { 'collectd::plugins::memcached':
  modules => {
    'config' => {
      'Host'  => '"127.0.0.1"',
      'Port'  => '"11211"'
    }
  }
}

See collectd-haproxy for configurable parameters.

####Class: collectd::plugin::haproxy

class { 'collectd::plugins::haproxy':
  version => 'v1.0.1',
  modules => {
    'config' => {
      'Socket'  => '/var/lib/haproxy/stats',
      'ProxyMonitor'  => ['backend',  'http-in',  'server1'],
      'ExcludeMetric' => ['session_limit', 'response_1xx'],
    }
  }
}

See collectd-memcached for configurable parameters and memcached configuration instructions.

####Class: collectd::plugin::mesos

class { 'collectd::plugins::mesos' :
  modules => {
    'mesos-master' => {
      'Cluster' => '"cluster-0"',
      'Instance' => '"master-0"',
      'Path' => '"/usr/sbin"',
      'Host' => '"localhost"',
      'Port' => '5050',
      'Verbose' => 'false',
    }
  }
}

See collectd-mesos for configurable parameters.

####Class: collectd::plugin::mongodb

class { 'collectd::plugins::mongodb' :
  modules => {
    'module1' => {
      'Host' => '"localhost"',
      'Port' => '"27017"',
      'User' => '"collectd"',
      'Password' => '"password"',
      'Database' => '"db1"',
    },
    'module2' => {
      'Host' => '"localhost"',
      'Port' => '"27017"',
      'Database' => '"test"',
    }
  }
}

See collectd-mongodb for configurable parameters.

####Class: collectd::plugin::mysql

class { 'collectd::plugins::mysql' :
  modules => {
    'mydb_plugin_instance' => {
      'Host' => '"localhost"',
      'User' => '"admin"',
      'Password' => '"root"',
      'Database' => '"mydb"',
      'Socket' => '"/var/run/mysqld/mysqld.sock"'
    }
  }
}

See collectd.conf for configurable parameters. The sample output file generated would look like 10-mysql.conf

####Class: collectd::plugin::nginx

class { 'collectd::plugins::nginx':
  modules => {
    'config' => {
      'URL'  => '"http://localhost:80/nginx_status"',
    }
  }
}

See collectd-nginx for configurable parameters and nginx configuration instructions.

####Class: collectd::plugin::postgresql

class { 'collectd::plugins::postgresql' :
  modules => {
    'database1' => {
      'Host' => '"127.0.0.1"',
      'User' => '"postgres"',
      'Password' => '"password"',
      'queries' => [
        'custom_deadlocks',
        'backends',
        'transactions',
        'queries',
        'queries_by_table',
        'query_plans',
        'table_states',
        'query_plans_by_table',
        'table_states_by_table',
        'disk_io',
        'disk_io_by_table',
        'disk_usage',
      ]
    }
  }
}

See collectd-postgresql for configurable parameters.

####Class: collectd::plugin::rabbitmq

class { 'collectd::plugins::rabbitmq' :
  modules => {
    'rabbitmq-1' => {
      'Username' => '"guest"',
      'Password' => '"guest"',
      'Host' => '"localhost"',
      'Port' => '"15672"',
      'CollectChannels' => true,
      'CollectConnections' => true,
      'CollectExchanges' => true,
      'CollectNodes' => true,
      'CollectQueues' => true,
      'FieldLength' => '1024'
    }
  }
}

See collectd-rabbitmq for configurable parameters. The sample output file generated would look like 10-rabbitmq.conf. Currently, the plugin only monitors one rabbitmq instance, so you should include only one module in the above class arguments.

####Class: collectd::plugin::redis

class { 'collectd::plugins::redis' :
  modules => {
    'redis_info' => {
      'Host' => '"localhost"',
      'Port' => 6379,
      'Verbose' => 'false',
      'Redis_uptime_in_seconds' => '"gauge"',
      'Redis_used_cpu_sys' => '"counter"',
      'Redis_used_cpu_user' => '"counter"',
      'Redis_used_cpu_sys_children' => '"counter"',
      'Redis_used_cpu_user_children' => '"counter"',
      'Redis_uptime_in_days' => '"gauge"',
      'Redis_lru_clock' => '"counter"',
      'Redis_connected_clients' => '"gauge"',
      'Redis_connected_slaves' => '"gauge"',
      'Redis_client_longest_output_list' => '"gauge"',
      'Redis_client_biggest_input_buf' => '"gauge"',
      'Redis_blocked_clients' => '"gauge"',
      'Redis_expired_keys' => '"counter"',
      'Redis_evicted_keys' => '"counter"',
      'Redis_rejected_connections' => '"counter"',
      'Redis_used_memory' => '"bytes"',
      'Redis_used_memory_rss' => '"bytes"',
      'Redis_used_memory_peak' => '"bytes"',
      'Redis_used_memory_lua' => '"bytes"',
      'Redis_mem_fragmentation_ratio' => '"gauge"',
      'Redis_changes_since_last_save' => '"gauge"',
      'Redis_instantaneous_ops_per_sec' => '"gauge"',
      'Redis_rdb_bgsave_in_progress' => '"gauge"',
      'Redis_total_connections_received' => '"counter"',
      'Redis_total_commands_processed' => '"counter"',
      'Redis_total_net_input_bytes' => '"counter"',
      'Redis_total_net_output_bytes' => '"counter"',
      'Redis_keyspace_hits' => '"derive"',
      'Redis_keyspace_misses' => '"derive"',
      'Redis_latest_fork_usec' => '"gauge"',
      'Redis_connected_slaves' => '"gauge"',
      'Redis_repl_backlog_first_byte_offset' => '"gauge"',
      'Redis_master_repl_offset' => '"gauge"',
    }
  }
}

See redis-collectd-plugin for configurable parameters. The sample output file generated would look like 10-redis_master.conf

####Class: collectd::plugin::varnish

class { 'collectd::plugins::varnish' :
  modules => {
    'instance' => {
      'CollectCache' => true,
      'CollectConnections' => true,
      'CollectBackend' => true,
      'CollectSHM' => true,
      'CollectESI' => true,
      'CollectFetch' => true,
      'CollectHCB' => true,
      'CollectSMA' => true,
      'CollectSMS' => true,
      'CollectSM' => true,
      'CollectTotals' => true,
      'CollectWorkers' => true,
      'CollectUptime' => true,
      'CollectVCL' => true,
      'CollectStruct' => true,
      'CollectObjects' => true,
      'CollectSession' => true,
      'CollectVSM' =>true
    }
  }
}

See collectd-varnish for configurable parameters.

####Class: collectd::plugin::vmstat

class { 'collectd::plugins::vmstat':
  modules => {
    'vmstat_collectd' => {
        'Path' => '"/usr/bin/vmstat"',
        'Verbose' => false,
        'Include' => '["r", "b", "swpd", "free", "buff", "cache", "inact", "active", "si", "so", "bi", "bo", "in", "cs", "us", "sy", "id", "wa", "st"]'
    }
  }
}

See collectd-vmstat for configurable parameters.

####Class: collectd::plugin::zookeeper

class { 'collectd::plugins::zookeeper' :
  modules => {
    'module' => {
      'Hosts' => '"localhost"',
      'Port' => '2181',
    }
  }
}

See collectd-zookeeper for configurable parameters.

Supported Platforms

Currently, the supported platforms for this module are:

1. Ubuntu 12.04
2. Ubuntu 14.04
3. Ubuntu 15.04
4. Ubuntu 16.04
5. Ubuntu 18.04
6. CentOS 6
7. CentOS 7
8. RHEL 6
9. RHEL 7
10. Amazon Linux 2014.09
11. Amazon Linux 2015.03
12. Amazon Linux 2015.09
13. Amazon Linux 2016.03
14. Amazon Linux 2016.09
15. Amazon Linux 2017.03
16. Amazon Linux 2017.09
17. Amazon Linux 2017.12
18. Amazon Linux 2018.03
19. Debian GNU/Linux 7 (wheezy)
20. Debian GNU/Linux 8 (jessie)
21. Debian GNU/Linux 9 (stretch)

Release 0.1.18

• Support for Ubuntu 18.04 and Amazon Linux 2017.12 and 2018.03
• Update minimum version of puppetlabs/apt to 2.3.0

Release 0.1.17

• Add HAProxy plugin

Release 0.1.16

• Installing dirmngr on Debian Stretch so GPG keys get imported right

Release 0.1.15

• Supporting Debian Stretch

Release 0.1.14

• Fixing typo in collectd disk plugin manifest

Release 0.1.13

• Adding support for Amazon Linux 2016.09, 2017.03, and 2017.09

Release 0.1.12

• Fix for plugin modules configuration

Release 0.1.11

• Add ability to disable individual default plugins
• Add df plugin as custom configurable
• Add disk plugin as custom configurable

Release 0.1.10

• Allow setting github repo revision to pull from in get_from_github
• Standardize all plugin templates to use modules parameter
• Introduce ability to set template location
• Introduce ability to set package version and use

Release 0.1.9

• Fix for vmstat plugin to support RHEL

Release 0.1.7

• Add ability to filter on plugin-specific metrics
• Add vmstat plugin
• Change nginx plugin usage
• Update docker plugin to v1.0.0 release
• Update elatsicsearch plugin and configuration to 1.2.0 release.

Release 0.1.6

• Add ability to configure plugin-specific reporting intervals
• Add iostat plugin
• Add memcached plugin
• Add varnish plugin
• Update signalfx metadata plugin to support optionally reporting cpu utilization on a per-core basis

Release 0.1.5

• Add apache plugin
• Add cassandra plugin
• Add docker plugin
• Update elasticsearch plugin
• Add java (supporting) plugin
• Add kafka plugin
• Add mesos (master) plugin
• Add mongodb plugin
• Add nginx plugin
• Add postgresql plugin
• Update redis plugin
• Add zookeeper plugin

Release 0.1.4

• Support Amazon Linux 2016.03

Release 0.1.3

• Fix file permissions
• Revert network unreachable error fix, its not module's error

Release 0.1.2

• Fix network unreachable error on non AWS puppetmaster

Release 0.1.1

• Update version of stdlib to support apt

Release 0.1.0

• Merge three modules into one module
• Add aggregation, elasticsearch, mysql, rabbitmq, redis plugin
• Configurable parameters to each plugin