subcategory |
---|
Deployment |
-> Note Initialize provider with alias = "mws"
, host = "https://accounts.cloud.databricks.com"
and use provider = databricks.mws
This resource to configure KMS keys for new workspaces within AWS or GCP. This is to support the following features:
- Customer-managed keys for managed services: Encrypt the workspace’s managed services data in the control plane, including notebooks, secrets, Databricks SQL queries, and Databricks SQL query history with a CMK.
- Customer-managed keys for workspace storage: Encrypt the workspace's root S3 bucket and clusters' EBS volumes with a CMK.
Please follow this complete runnable example with new VPC and new workspace setup. Please pay special attention to the fact that there you have two different instances of a databricks provider - one for deploying workspaces (with host="https://accounts.cloud.databricks.com/"
) and another for the workspace you've created with databricks_mws_workspaces resource. If you want both creation of workspaces & clusters within workspace within the same terraform module (essentially same directory), you should use the provider aliasing feature of Terraform. We strongly recommend having one Terraform module for creation of workspace + PAT token and the rest in different modules.
-> Note If you've used the resource before, please add use_cases = ["MANAGED_SERVICES"]
to keep the previous behaviour.
You must configure this during workspace creation
variable "databricks_account_id" {
description = "Account Id that could be found in the top right corner of https://accounts.cloud.databricks.com/"
}
data "aws_caller_identity" "current" {}
data "aws_iam_policy_document" "databricks_managed_services_cmk" {
version = "2012-10-17"
statement {
sid = "Enable IAM User Permissions"
effect = "Allow"
principals {
type = "AWS"
identifiers = [data.aws_caller_identity.current.account_id]
}
actions = ["kms:*"]
resources = ["*"]
}
statement {
sid = "Allow Databricks to use KMS key for control plane managed services"
effect = "Allow"
principals {
type = "AWS"
identifiers = ["arn:aws:iam::414351767826:root"]
}
actions = [
"kms:Encrypt",
"kms:Decrypt"
]
resources = ["*"]
}
}
resource "aws_kms_key" "managed_services_customer_managed_key" {
policy = data.aws_iam_policy_document.databricks_managed_services_cmk.json
}
resource "aws_kms_alias" "managed_services_customer_managed_key_alias" {
name = "alias/managed-services-customer-managed-key-alias"
target_key_id = aws_kms_key.managed_services_customer_managed_key.key_id
}
resource "databricks_mws_customer_managed_keys" "managed_services" {
account_id = var.databricks_account_id
aws_key_info {
key_arn = aws_kms_key.managed_services_customer_managed_key.arn
key_alias = aws_kms_alias.managed_services_customer_managed_key_alias.name
}
use_cases = ["MANAGED_SERVICES"]
}
# supply databricks_mws_customer_managed_keys.managed_services.customer_managed_key_id as managed_services_customer_managed_key_id for databricks_mws_workspaces
variable "databricks_account_id" {
description = "Account Id that could be found in the top right corner of https://accounts.gcp.databricks.com/"
}
variable "cmek_resource_id" {
description = "Id of a google_kms_crypto_key"
}
resource "databricks_mws_customer_managed_keys" "managed_services" {
account_id = var.databricks_account_id
gcp_key_info {
kms_key_id = var.cmek_resource_id
}
use_cases = ["MANAGED_SERVICES"]
}
# supply databricks_mws_customer_managed_keys.managed_services.customer_managed_key_id as managed_services_customer_managed_key_id for databricks_mws_workspaces
variable "databricks_account_id" {
description = "Account Id that could be found in the top right corner of https://accounts.cloud.databricks.com/"
}
variable "databricks_cross_account_role" {
description = "AWS ARN for the Databricks cross account role"
}
data "aws_iam_policy_document" "databricks_storage_cmk" {
version = "2012-10-17"
statement {
sid = "Enable IAM User Permissions"
effect = "Allow"
principals {
type = "AWS"
identifiers = [data.aws_caller_identity.current.account_id]
}
actions = ["kms:*"]
resources = ["*"]
}
statement {
sid = "Allow Databricks to use KMS key for DBFS"
effect = "Allow"
principals {
type = "AWS"
identifiers = ["arn:aws:iam::414351767826:root"]
}
actions = [
"kms:Encrypt",
"kms:Decrypt",
"kms:ReEncrypt*",
"kms:GenerateDataKey*",
"kms:DescribeKey"
]
resources = ["*"]
}
statement {
sid = "Allow Databricks to use KMS key for DBFS (Grants)"
effect = "Allow"
principals {
type = "AWS"
identifiers = ["arn:aws:iam::414351767826:root"]
}
actions = [
"kms:CreateGrant",
"kms:ListGrants",
"kms:RevokeGrant"
]
resources = ["*"]
condition {
test = "Bool"
variable = "kms:GrantIsForAWSResource"
values = ["true"]
}
}
statement {
sid = "Allow Databricks to use KMS key for EBS"
effect = "Allow"
principals {
type = "AWS"
identifiers = [var.databricks_cross_account_role]
}
actions = [
"kms:Decrypt",
"kms:GenerateDataKey*",
"kms:CreateGrant",
"kms:DescribeKey"
]
resources = ["*"]
condition {
test = "ForAnyValue:StringLike"
variable = "kms:ViaService"
values = ["ec2.*.amazonaws.com"]
}
}
}
resource "aws_kms_key" "storage_customer_managed_key" {
policy = data.aws_iam_policy_document.databricks_storage_cmk.json
}
resource "aws_kms_alias" "storage_customer_managed_key_alias" {
name = "alias/storage-customer-managed-key-alias"
target_key_id = aws_kms_key.storage_customer_managed_key.key_id
}
resource "databricks_mws_customer_managed_keys" "storage" {
account_id = var.databricks_account_id
aws_key_info {
key_arn = aws_kms_key.storage_customer_managed_key.arn
key_alias = aws_kms_alias.storage_customer_managed_key_alias.name
}
use_cases = ["STORAGE"]
}
# supply databricks_mws_customer_managed_keys.storage.customer_managed_key_id as storage_customer_managed_key_id for databricks_mws_workspaces
variable "databricks_account_id" {
description = "Account Id that could be found in the top right corner of https://accounts.gcp.databricks.com/"
}
variable "cmek_resource_id" {
description = "Id of a google_kms_crypto_key"
}
resource "databricks_mws_customer_managed_keys" "storage" {
account_id = var.databricks_account_id
gcp_key_info {
kms_key_id = var.cmek_resource_id
}
use_cases = ["STORAGE"]
}
# supply databricks_mws_customer_managed_keys.storage.customer_managed_key_id as storage_customer_managed_key_id for databricks_mws_workspaces
The following arguments are required:
aws_key_info
- This field is a block and is documented below. This conflicts withgcp_key_info
gcp_key_info
- This field is a block and is documented below. This conflicts withaws_key_info
account_id
- Account Id that could be found in the top right corner of Accounts Consoleuse_cases
- (since v0.3.4) List of use cases for which this key will be used. If you've used the resource before, please adduse_cases = ["MANAGED_SERVICES"]
to keep the previous behaviour. Possible values are:MANAGED_SERVICES
- for encryption of the workspace objects (notebooks, secrets) that are stored in the control planeSTORAGE
- for encryption of the DBFS Storage & Cluster EBS Volumes
key_arn
- The AWS KMS key's Amazon Resource Name (ARN).key_alias
- (Optional) The AWS KMS key alias.key_region
- (Optional) (Computed) The AWS region in which KMS key is deployed to. This is not required.
kms_key_id
- The GCP KMS key's resource name.
In addition to all arguments above, the following attributes are exported:
id
- Canonical unique identifier for the mws customer managed keys.customer_managed_key_id
- (String) ID of the encryption key configuration object.creation_time
- (Integer) Time in epoch milliseconds when the customer key was created.
-> Note Importing this resource is not currently supported.
The following resources are used in the same context:
- Provisioning Databricks on AWS guide.
- databricks_mws_credentials to configure the cross-account role for creation of new workspaces within AWS.
- databricks_mws_log_delivery to configure delivery of billable usage logs and audit logs.
- databricks_mws_networks to configure VPC & subnets for new workspaces within AWS.
- databricks_mws_storage_configurations to configure root bucket new workspaces within AWS.
- databricks_mws_workspaces to set up AWS and GCP workspaces.