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):

Parameter	Description
signalfx_api_token	Your SignalFx API Token
dimension_list	Set custom dimensions on all of the metrics that collectd sends to SignalFx. For example, you can use a custom dimension to indicate that one of your servers is running Kafka by including it in the hash map as follows: dimension_list => {"serverType" => "kafka"}
aws_integration	Controls AWS metadata syncing to SignalFx. Default is true.
signalfx_api_endpoint	The API endpoint to post your metrics. This will be useful if you are using a proxy.
ensure_signalfx_collectd_version	Ensures the collectd version on the system. Accepted values are of ensure from Puppet.
signalfx_collectd_repo_source	The source of the collectd repository from SignalFx. This will be useful when you mirror a SignalFx repository. Valid on Ubuntu and Debian systems.
signalfx_plugin_repo_source	The source of the signalfx-collectd-plugin repository from SignalFx. This will be useful when you mirror a SignalFx repository. Valid on Ubuntu and Debian systems.
fqdnlookup	Fqdnlookup of the collectd.conf file
hostname	Hostname to be used if fqdnlookup is true, default value would be the hostname from Puppet Facter.
interval	Interval of the collectd.conf file
timeout	Timeout of the collectd.conf file
read_threads	ReadThreads of the collectd.conf file
write_queue_limit_high	WriteQueueLimitHigh of the collectd.conf file
write_queue_limit_low	WriteQueueLimitLow of the collectd.conf file
collect_internal_stats	CollectInternalStats of the collectd.conf file
log_file	The location of log file to be used by collectd
log_level	The log level to be used by collectd
use_default_cpu_plugin	Cpu default plugin option. Set to false to disable default cpu plugin
use_default_cpufreq_plugin	Cpufreq default plugin option. Set to false to disable default cpufreq plugin
use_default_df_plugin	Df default plugin option. Set to false to disable default df plugin
use_default_disk_plugin	Disk default plugin option. Set to false to disable default disk plugin
use_default_interface_plugin	Interface default plugin option. Set to false to disable default interface plugin
use_default_load_plugin	Load default plugin option. Set to false to disable default load plugin
use_default_memory_plugin	Memory default plugin option. Set to false to disable default memory plugin
use_default_protocols_plugin	Protocols default plugin option. Set to false to disable default protocols plugin
use_default_vmem_plugin	Vmem default plugin option. Set to false to disable default vmem plugin
use_default_uptime_plugin	Uptime default plugin option. Set to false to disable default uptime plugin
write_http_timeout	Timeout option of write_http plugin
write_http_buffersize	BufferSize option of write_http plugin
write_http_flush_interval	FlushInterval option of write_http plugin
write_http_log_http_error	LogHttpError option of write_http plugin
ensure_signalfx_plugin_version	Ensures the signalfx-collectd-plugin version on the system. Accepted values are of ensure from Puppet.
signalfx_plugin_log_traces	LogTraces of signalfx-collectd-plugin
signalfx_plugin_interactive	Interactive option of signalfx-collectd-plugin
signalfx_plugin_notifications	Notifications option of signalfx-collectd-plugin
signalfx_plugin_notify_level	NotifyLevel option of signalfx-collectd-plugin
signalfx_plugin_dpm	DPM option of signalfx-collectd-plugin
signalfx_plugin_utilization	Utilization option of signalfx-collectd-plugin
signalfx_plugin_cpu_utilization	CPU Utilization option of signalfx-collectd-plugin
signalfx_plugin_cpu_utilization_per_core	CPU Utilization for individual cores option of signalfx-collectd-plugin

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. Kafka
8. Iostat
9. Memcached
10. Mesos
11. MongoDB
12. MySQL
13. Nginx
14. Postgresql
15. RabbitMQ
16. Redis
17. Varnish
18. Vmstat
19. 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-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. CentOS 6
6. CentOS 7
7. RHEL 6
8. RHEL 7
9. Amazon Linux 2014.09
10. Amazon Linux 2015.03
11. Amazon Linux 2015.09
12. Amazon Linux 2016.03
13. Debian GNU/Linux 7 (wheezy)
14. Debian GNU/Linux 8 (jessie)