agent

This library is a tool for writing netdata plugins.

We strongly believe that custom plugins are very important and they must be easy to write.

Definitions:

orchestrator
plugin orchestrators are external plugins that do not collect any data by themselves. Instead they support data collection modules written in the language of the orchestrator. Usually the orchestrator provides a higher level abstraction, making it ideal for writing new data collection modules with the minimum of code.
plugin
plugin is a set of data collection modules.
module
module is a data collector. It collects, processes and returns processed data to the orchestrator.
job
job is a module instance with specific settings.

Package provides:

CLI parser
plugin orchestrator (loads configurations, creates and serves jobs)

You are responsible only for creating modules.

Custom plugin example#

Yep! So easy!

How to write a Module#

Module is responsible for charts creating and data collecting. Implement Module interface and that is it.

type Module interface {
    // Init does initialization.
    // If it returns false, the job will be disabled.
    Init() bool

    // Check is called after Init.
    // If it returns false, the job will be disabled.
    Check() bool

    // Charts returns the chart definition.
    // Make sure not to share returned instance.
    Charts() *Charts

    // Collect collects metrics.
    Collect() map[string]int64

    // SetLogger sets logger.
    SetLogger(l *logger.Logger)

    // Cleanup performs cleanup if needed.
    Cleanup()
}

// Base is a helper struct. All modules should embed this struct.
type Base struct {
    *logger.Logger
}

// SetLogger sets logger.
func (b *Base) SetLogger(l *logger.Logger) { b.Logger = l }

How to write a Plugin#

Since plugin is a set of modules all you need is:

write module(s)
add module(s) to the plugins registry
start the plugin

How to integrate your plugin into Netdata#

Three simple steps:

move the plugin to the plugins.d dir.
add plugin configuration file to the etc/netdata/ dir.
add modules configuration files to the etc/netdata/<DIR_NAME>/ dir.

Congratulations!

Configurations#

Configurations are written in YAML.

plugin configuration:

# Enable/disable the whole plugin.
enabled: yes

# Default enable/disable value for all modules.
default_run: yes

# Maximum number of used CPUs. Zero means no limit.
max_procs: 0

# Enable/disable specific plugin module
modules:
#  module_name1: yes
#  module_name2: yes

module configuration

# [ GLOBAL ]
update_every: 1
autodetection_retry: 0

# [ JOBS ]
jobs:
  - name: job1
    param1: value1
    param2: value2

  - name: job2
    param1: value1
    param2: value2

Plugin uses yaml.Unmarshal to add configuration parameters to the module. Please use yaml tags!

Debug#

Plugin CLI:

Usage:
  plugin [OPTIONS] [update every]

Application Options:
  -d, --debug    debug mode
  -m, --modules= modules name (default: all)
  -c, --config=  config dir

Help Options:
  -h, --help     Show this help message

Specific module debug:

# become user netdata
sudo su -s /bin/bash netdata

# run plugin in debug mode
./<plugin_name> -d -m <module_name>

Change <plugin_name> to your plugin name and <module_name> to the module name you want to debug.