gcp_compute – Google Cloud Compute Engine inventory source

Synopsis

  • Get inventory hosts from Google Cloud Platform GCE.
  • Uses a YAML configuration file that ends with gcp_compute.(yml|yaml) or gcp.(yml|yaml).

Requirements

The below requirements are needed on the local master node that executes this inventory.

  • requests >= 2.18.4
  • google-auth >= 1.3.0

Parameters

Parameter Choices/Defaults Configuration Comments
auth_kind
- / required
    Choices:
  • application
  • serviceaccount
  • machineaccount
env:GCP_AUTH_KIND
The type of credential used.
cache
boolean
    Choices:
  • no ←
  • yes
ini entries:

[inventory]
cache = no

env:ANSIBLE_INVENTORY_CACHE
Toggle to enable/disable the caching of the inventory's source data, requires a cache plugin setup to work.
cache_connection
string
ini entries:

[defaults]
fact_caching_connection = VALUE

[inventory]
cache_connection = VALUE

env:ANSIBLE_CACHE_PLUGIN_CONNECTION
env:ANSIBLE_INVENTORY_CACHE_CONNECTION
Cache connection data or path, read cache plugin documentation for specifics.
cache_plugin
string
Default:
"memory"
ini entries:

[defaults]
fact_caching = memory

[inventory]
cache_plugin = memory

env:ANSIBLE_CACHE_PLUGIN
env:ANSIBLE_INVENTORY_CACHE_PLUGIN
Cache plugin to use for the inventory's source data.
cache_prefix
-
Default:
"ansible_inventory_"
ini entries:

[default]
fact_caching_prefix = ansible_inventory_

[inventory]
cache_prefix = ansible_inventory_

env:ANSIBLE_CACHE_PLUGIN_PREFIX
env:ANSIBLE_INVENTORY_CACHE_PLUGIN_PREFIX
Prefix to use for cache plugin files/tables
cache_timeout
integer
Default:
3600
ini entries:

[defaults]
fact_caching_timeout = 3600

[inventory]
cache_timeout = 3600

env:ANSIBLE_CACHE_PLUGIN_TIMEOUT
env:ANSIBLE_INVENTORY_CACHE_TIMEOUT
Cache duration in seconds
compose
dictionary
Default:
{}
Create vars from jinja2 expressions.
filters
list
A list of filter value pairs. Available filters are listed here https://cloud.google.com/compute/docs/reference/rest/v1/instances/aggregatedList. Each additional filter in the list will act be added as an AND condition (filter1 and filter2)
groups
dictionary
Default:
{}
Add hosts to group based on Jinja2 conditionals.
hostnames
list
Default:
["public_ip", "private_ip", "name"]
A list of options that describe the ordering for which hostnames should be assigned. Currently supported hostnames are 'public_ip', 'private_ip', or 'name'.
keyed_groups
list
Default:
[]
Add hosts to group based on the values of a variable.
plugin
- / required
    Choices:
  • gcp_compute
token that ensures this is a source file for the 'gcp_compute' plugin.
projects
list / required
A list of projects in which to describe GCE instances.
retrieve_image_info
boolean
added in 2.8
    Choices:
  • no ←
  • yes
Populate the image host fact for the instances returned with the GCP image name
By default this plugin does not attempt to resolve the boot image of an instance to the image name cataloged in GCP because of the performance overhead of the task.
Unless this option is enabled, the image host variable will be null
scopes
list
Default:
["https://www.googleapis.com/auth/compute"]
env:GCP_SCOPES
list of authentication scopes
service_account_contents
string
added in 2.8.2
env:GCP_SERVICE_ACCOUNT_CONTENTS
A string representing the contents of a Service Account JSON file. This should not be passed in as a dictionary, but a string that has the exact contents of a service account json file (valid JSON).
service_account_email
-
env:GCP_SERVICE_ACCOUNT_EMAIL
An optional service account email address if machineaccount is selected and the user does not wish to use the default email.
service_account_file
path
env:GCP_SERVICE_ACCOUNT_FILE
env:GCE_CREDENTIALS_FILE_PATH
The path of a Service Account JSON file if serviceaccount is selected as type.
strict
boolean
    Choices:
  • no ←
  • yes
If yes make invalid entries a fatal error, otherwise skip and continue.
Since it is possible to use facts in the expressions they might not always be available and we ignore those errors by default.
use_contrib_script_compatible_sanitization
boolean
added in 2.8
    Choices:
  • no ←
  • yes
By default this plugin is using a general group name sanitization to create safe and usable group names for use in Ansible. This option allows you to override that, in efforts to allow migration from the old inventory script.
For this to work you should also turn off the TRANSFORM_INVALID_GROUP_CHARS setting, otherwise the core engine will just use the standard sanitization on top.
This is not the default as such names break certain functionality as not all characters are valid Python identifiers which group names end up being used as.
vars_prefix
-
Default:
""
prefix to apply to host variables, does not include facts nor params
zones
list
A list of regions in which to describe GCE instances. If none provided, it defaults to all zones available to a given project.

Examples

plugin: gcp_compute
zones: # populate inventory with instances in these regions
  - us-east1-a
projects:
  - gcp-prod-gke-100
  - gcp-cicd-101
filters:
  - machineType = n1-standard-1
  - scheduling.automaticRestart = true AND machineType = n1-standard-1
service_account_file: /tmp/service_account.json
auth_kind: serviceaccount
scopes:
 - 'https://www.googleapis.com/auth/cloud-platform'
 - 'https://www.googleapis.com/auth/compute.readonly'
keyed_groups:
  # Create groups from GCE labels
  - prefix: gcp
    key: labels
hostnames:
  # List host by name instead of the default public ip
  - name
compose:
  # Set an inventory parameter to use the Public IP address to connect to the host
  # For Private ip use "networkInterfaces[0].networkIP"
  ansible_host: networkInterfaces[0].accessConfigs[0].natIP

Status

Authors

  • UNKNOWN

Hint

If you notice any issues in this documentation, you can edit this document to improve it.

Hint

Configuration entries for each entry type have a low to high priority order. For example, a variable that is lower in the list will override a variable that is higher up.