influxdb

Table of Contents

1. Description
2. Setup - The basics of getting started with influxdb
   • What influxdb affects
   • Beginning with influxdb
3. Usage - Configuration options and additional functionality
4. Limitations - OS compatibility, etc.

Description

This module provides type and provider implementations to manage the resources of an InfluxDB 2.x instance. Because the InfluxDB 2.0 api provides an interface to these resources, the module is able to manage an InfluxDB server running either on the local machine or remotely.

Setup

What influxdb affects

The primary things this module provides are:

• Installation of InfluxDB repositories and packages
• Initial setup of the InfluxDB application
• Configuration and management of InfluxDB resources such as organizations, buckets, etc

The first two items are provided by the influxdb class and are restricted to an InfluxDB instance running on the local machine.

InfluxDB resources are managed by the various types and providers. Because we need to be able to enumerate and query resources on either a local or remote machine, the resources accept these parameters with the following defaults:

• host - fqdn
• port - 8086
• token_file - ~/.influxdb_token
• use_ssl - true
• token (optional)

Specifying a token in Sensitive[String] format is optional, but recommended. See Beggining with Influxdb for more info.

Note that you are not able to use multiple combinations of these options in a given catalog. Each provider class will set these values when first instantiated and will use the first value that it finds. Therefore, it is best to use resource defaults for these parameters in your manifest, e.g.

class my_profile::my_class(
  Sensitive[String] $my_token,
){
  Influxdb_bucket {
    token => $my_token,
  }
}

See Usage for more information about these use cases.

Beginning with InfluxDB

The easiest way to get started using this module is by including the influxdb class to install and perform initial setup of the application.

include influxdb

Doing so will:

• Install the influxdb2 package from either a repository or archive source.
• Configure and start the influxdb service
• Perform initial setup of the InfluxDB application, consisting of
  • An initial organization and bucket
  • An administrative token saved to ~/.influxdb_token by default

The type and provider code is able to use the token saved in this file, provided it is present on the node applying the catalog. However, it is recommended to specify the token via the influxdb::token parameter after initial setup.

Usage

Installation

As detailed in [Beginning with influxdb](#Beginning with influxdb), the influxdb class manages installation and initial setup of InfluxDB. The following aspects are managed by default:

• InfluxDB repository
• SSL
• Initial setup, including the initial organization and bucket resources
• Token with permissions to read and write Telegrafs and buckets within the initial organization

Note that the admin user and password can be set prior to initial setup, but cannot be managed afterwards. These must be changed manually using the influx cli.

For example, to use a different initial organization and bucket, set the parameters in hiera:

influxdb::initial_org: 'my_org'
influxdb::initial_bucket: 'my_bucket'

Or use a class-like declaration

class {'influxdb':
  initial_org    => 'my_org',
  initial_bucket => 'my_bucket',
}

Resource management

For managing InfluxDB resources, this module provides several types and providers that use the InfluxDB 2.0 api. As mentioned in What influxdb affects, the resources accept parameters to determine how to connect to the host which must be unique per resource type. For example, to create an organization and bucket and specify a token and non-standard port:

class my_profile::my_class(
  Sensitive[String] $token,
){

  influxdb_org {'my_org':
    ensure => present,
    token  => $token,
    port   => 1234,
  }

  influxdb_bucket {'my_bucket':
    ensure  => present,
    org     => 'my_org',
    labels  => ['my_label1', 'my_label2'],
    token  => $token,
    port   => 1234,
  }
}

Resource defaults are also a good option:

Influxdb_org {
  token => $token,
  port  => 1234,
}

Influxdb_bucket {
  token => $token,
  port  => 1234,
}

Note that the influxdb_bucket will produce a warning for each specified label that does not currently exist.

If InfluxDB is running locally and there is an admin token saved at ~/.influxdb_token, it will be used in API calls if the token parameter is unset. However, it is recommended to set the token in hiera as an eyaml-encrypted string. For example:

influxdb::token: '<eyaml_string>'
lookup_options:
   influxdb::token:
     convert_to: "Sensitive"

For more complex resource management, here is an example of:

• Looking up a list of buckets
• Creating a hash with ensure => present for each bucket
• Creating the bucket resources with a default org of myorg and retention policy of 30 days.

Hiera data:

profile::buckets:
  - 'bucket1'
  - 'bucket2'
  - 'bucket3'

Puppet code:

class my_profile::my_class{
  $buckets = lookup('profile::buckets')
  $bucket_hash = $buckets.reduce({}) |$memo, $bucket| {
    $tmp = $memo.merge({"$bucket" => { "ensure" => present } })
    $tmp
  }

  create_resources(
    influxdb_bucket,
    $bucket_hash,
    {
      'org'        => 'myorg',
      retention_rules => [{
        'type' => 'expire',
        'everySeconds' => 2592000,
        'shardGroupDurationSeconds' => 604800,
      }]
    }
  )

Limitations

This module is incompatible with InfluxDB 1.x. Migrating data from 1.x to 2.x must be done manually. For more information see here.