Skip to content

Be notified in the event of a new GCP Organization Policy being released.

License

Notifications You must be signed in to change notification settings

ScaleSec/gcp_org_policy_notifier

Repository files navigation

GCP Organization Policy Notifier

GitHub Super-Linter

Description

This is a 100% serverless tool that analyzes GCP Organization Policies for updates and then posts to a slack channel as well as twitter via our Twitter bot.

Process Flow

Organization Policy Notifier Flow

  1. A Cloud Scheduler job kicks off the comparison every hour and sends a base64 encoded message to Cloud Pub/Sub.
  2. Cloud Pub/Sub forwards the encoded message to a Cloud Function via a Pub/Sub Subscription.
  3. The Cloud Function receives the message, decodes it, and announces the beginning of the comparison.
  4. The Cloud Function searches for a pre-existing Organization Policy Constraint baseline file in a GCS bucket.
  • If a baseline file exists: it copies the file locally for comparison.
  • If a baseline file does not exist in the GCS bucket: The function creates a baseline based on the current available Organization Policy Constraints and then uploads to GCS for future comparisons.
  1. After the baseline is copied locally (if a baseline existed), the Cloud Function queries the available Organization Policy Constraints performs a comparison.
  • If there are updates, the new Organization Policy Constraint list that was generated becomes the new baseline and is updated to the GCS bucket for future comparisons. The Cloud Function then moves on to steps 6, 7, and 8.
  1. The Cloud Function creates a GitHub Pull Request with the new Organization Policies which can be found here.
  2. The Cloud Function posts to Twitter via the handle @gcporgpolicybot and includes the new constraints plus a link to the GitHub commit.
  3. In addition to a Twitter post, the Cloud Function will post in a Slack Channel alerting the participants of the new constraints.

Configuration

  1. Update or comment out the backend.tf file for the terraform state file.
terraform {
  backend "gcs" {
    bucket = "<insert_value_here>"
    prefix = "<insert_value_here>"
  }
}
  1. In order to keep secrets out of the Terraform state file, we recommend manually uploading the Slack Webhook into GCP Secrets Manager and then referencing that in the below terraform variables.

  2. Fill in the required values for the terraform.tfvars file. We recommend an isolated project for this solution as well as a separate GCS bucket for your policy file and the function's code file src.zip

project_id                       = ""
org_id                           = ""
secret_project                   = ""
name_prefix                      = ""
secret_slack_name                = ""
secret_token_name                = ""
secret_version                   = ""
twitter_consumer_key_name        = ""
twitter_consumer_key_secret_name = ""
twitter_access_token_name        = ""
twitter_access_token_secret_name = ""
policy_bucket_location.      = ""

Deployment

  1. Clone the repository locally:
git clone git@github.com:ScaleSec/gcp_org_policy_notifier.git
  1. Create your virtual environment:
python3 -m venv my_venv
  1. Activate environment and install dependencies:
source my_venv/bin/activate
pip install -r src/requirements.txt
  1. Deploy via terraform:
terraform init
terraform plan 
terraform apply

Inputs

Name Description Type Default Required
bucket_force_destroy When deleting the GCS bucket containing the cloud function, delete all objects in the bucket first. bool true no
file_location Location to store the org policy file in the Cloud Function. Needs to be in /tmp/. string "/tmp/policies.txt" no
function_available_memory_mb The amount of memory in megabytes allotted for the function to use. number 2048 no
function_description The description of the function. string "Compares Org Policies and alerts users." no
function_entry_point The name of a method in the function source which will be invoked when the function is executed. string "announce_kickoff" no
function_event_trigger_failure_policy_retry A toggle to determine if the function should be retried on failure. bool false no
function_perms The Cloud Function custom IAM role permissions. Must be a list. list
[
"secretmanager.secrets.get",
"secretmanager.versions.get",
"secretmanager.versions.access",
"orgpolicy.policy.get",
"resourcemanager.projects.get",
"resourcemanager.projects.list",
"storage.objects.create",
"storage.objects.get",
"storage.objects.update",
"storage.objects.delete",
"storage.objects.list"
]
no
function_runtime The runtime in which the function will be executed. string "python37" no
function_source_directory The contents of this directory will be archived and used as the function source. string "./src" no
function_timeout_s The amount of time in seconds allotted for the execution of the function. number 60 no
job_description The description of the Cloud Scheduler. string "Starts Organization Policies check." no
job_schedule The job frequency, in cron syntax. The default is every hour. string "0 * * * *" no
message_data The data to send in the topic message. string "U3RhcnRpbmcgQ29tcGFyaXNvbg==" no
name_prefix The prefixed used to name resources string n/a yes
org_id The GCP Org ID to assign permissions to. any n/a yes
policy_file The name of the Org policy file in the GCS bucket. string "policies.txt" no
project_id The ID of the project where the resources will be created. string n/a yes
region The region in which resources will be applied. string "us-central1" no
scheduler_job An existing Cloud Scheduler job instance. object({ name = string }) null no
secret_project The GCP project the Slack Webhook is stored. any n/a yes
secret_slack_name The name of the Slack Webhook secret in GCP. any n/a yes
secret_token_name The name of the GitHub token secret in GCP. any n/a yes
secret_version The version of the Slack Webhook secret in GCP. Leave as an empty string to use 'latest' string "latest" no
time_zone The timezone to use in scheduler. string "America/Detroit" no
twitter_access_token_name The name of the Twitter Access Token secret in GCP. any n/a yes
twitter_access_token_secret_name The name of the Twitter Access Token Secret secret in GCP. any n/a yes
twitter_consumer_key_name The name of the Twitter Consumer Key secret in GCP. any n/a yes
twitter_consumer_key_secret_name The name of the Twitter Consumer Key Secret secret in GCP. any n/a yes

Feedback

Feedback is welcome and encouraged via a GitHub issue. Please open an issue for any bugs, feature requests, or general improvements you would like to see. Thank you in advance!

About

Be notified in the event of a new GCP Organization Policy being released.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published