DEV Community

Cover image for How to Backup the Terraform states from Terraform Cloud workspaces.
Pipo Sanfilippo
Pipo Sanfilippo

Posted on

How to Backup the Terraform states from Terraform Cloud workspaces.

This tool is the result of a no so sad story as it should have been. We had a backup.
Until today, Terraform Cloud does not provide a mechanism to backup and restore the terraform states of the workspaces. This tool is the first version of a lambda that every time a workspace changes in the terraform state, the Terraform state will save in an S3 Bucket. So in case of disaster, the Terraform state can be pushed to the workspace.

Let us begin.
This tool is a serverless and event-driven solution. We are going to use:
Lambda repository
API Gateway
Terraform Cloud Workspace Notifications

The workflow of the process is:
The terraform state change in terraform cloud because a new plan was applied or that a plan has been applied and ended in an errored state. In any case, the terraform state has changed so that we will save this new state.
This change of state triggers a notification in Terraform Cloud to a generic endpoint. This generic endpoint is an API Gateway endpoint that has a lambda function integrated.

The lambda function receives the payload from the terraform cloud notification, and this payload looks like this:

   "resource": "/{proxy+}",
     "path": "/hello/world",
     "httpMethod": "POST",
     "headers": {
        "Accept": "*/*",
        "Accept-Encoding": "gzip, deflate",
        "cache-control": "no-cache",
        "CloudFront-Forwarded-Proto": "https",
        "CloudFront-Is-Desktop-Viewer": "true",
        "CloudFront-Is-Mobile-Viewer": "false",
        "CloudFront-Is-SmartTV-Viewer": "false",
        "CloudFront-Is-Tablet-Viewer": "false",
        "CloudFront-Viewer-Country": "US",
        "Content-Type": "application/json",
        "headerName": "headerValue",
        "Host": "",
        "Postman-Token": "9f583ef0-ed83-4a38-aef3-eb9ce3f7a57f",
        "User-Agent": "PostmanRuntime/2.4.5",
        "Via": "1.1 (CloudFront)",
        "X-Amz-Cf-Id": "pn-PWIJc6thYnZm5P0NMgOUglL1DYtl0gdeJky8tqsg8iS_sgsKD1A==",
        "X-Forwarded-For": ",",
        "X-Forwarded-Port": "443",
        "X-Forwarded-Proto": "https"
    "multiValueHeaders": {
        "Accept": ["*/*"],
        "Accept-Encoding": ["gzip, deflate"],
        "cache-control": ["no-cache"],
        "CloudFront-Forwarded-Proto": ["https"],
        "CloudFront-Is-Desktop-Viewer": ["true"],
        "CloudFront-Is-Mobile-Viewer": ["false"],
        "CloudFront-Is-SmartTV-Viewer": ["false"],
        "CloudFront-Is-Tablet-Viewer": ["false"],
        "CloudFront-Viewer-Country": ["US"],
        "Content-Type": ["application/json"],
        "headerName": ["headerValue"],
        "Host": [""],
        "Postman-Token": ["9f583ef0-ed83-4a38-aef3-eb9ce3f7a57f"],
        "User-Agent": ["PostmanRuntime/2.4.5"],
        "Via": ["1.1 (CloudFront)"],
        "X-Amz-Cf-Id": ["pn-PWIJc6thYnZm5P0NMgOUglL1DYtl0gdeJky8tqsg8iS_sgsKD1A=="],
        "X-Forwarded-For": [","],
        "X-Forwarded-Port": ["443"],
        "X-Forwarded-Proto": ["https"]
   "queryStringParameters": {
      "name": "me"
    "multiValueQueryStringParameters": {
        "name": ["me"]
   "pathParameters": {
      "proxy": "hello/world"
   "stageVariables": {
      "stageVariableName": "stageVariableValue"
   "requestContext": {
      "accountId": "12345678912",
      "resourceId": "roq9wj",
      "stage": "testStage",
      "domainName": "",
      "domainPrefix": "y0ne18dixk",
      "requestId": "deef4878-7910-11e6-8f14-25afc3e9ae33",
      "protocol": "HTTP/1.1",
      "identity": {
         "cognitoIdentityPoolId": "theCognitoIdentityPoolId",
         "accountId": "theAccountId",
         "cognitoIdentityId": "theCognitoIdentityId",
         "caller": "theCaller",
            "apiKey": "theApiKey",
            "apiKeyId": "theApiKeyId",
            "accessKey": "ANEXAMPLEOFACCESSKEY",
         "sourceIp": "",
         "cognitoAuthenticationType": "theCognitoAuthenticationType",
         "cognitoAuthenticationProvider": "theCognitoAuthenticationProvider",
         "userArn": "theUserArn",
         "userAgent": "PostmanRuntime/2.4.5",
         "user": "theUser"
      "authorizer": {
         "principalId": "admin",
         "clientId": 1,
         "clientName": "Exata"
      "resourcePath": "/{proxy+}",
      "httpMethod": "POST",
      "requestTime": "15/May/2020:06:01:09 +0000",
      "requestTimeEpoch": 1589522469693,
      "apiId": "gy415nuibc"
   "body": "{\n  \"payload_version\":1,\n  \"notification_configuration_id\":\"nc-vxvQ5SpJJ23oAWLk\",\n  \"run_url\":\"\",\n  \"run_id\":\"run-uyeKDPGKNpD9PxdC\",\n  \"run_message\":\"Queued manually using Terraform\",\n  \"run_created_at\":\"2021-05-30T16:18:00.000Z\",\n  \"run_created_by\":\"mnsanfilippo\",\n  \"workspace_id\":\"ws-WzbRhM4ojZXjRvoP\",\n  \"workspace_name\":\"mnsanfilippo\",\n  \"organization_name\":\"mnsanfilippo\",\n  \"notifications\":[\n    {\n      \"message\":\"Run Errored\",\n      \"trigger\":\"run:errored\",\n      \"run_status\":\"errored\",\n      \"run_updated_at\":\"2021-05-30T16:19:14.000Z\",\n      \"run_updated_by\":null\n    }\n  ]\n}"
Enter fullscreen mode Exit fullscreen mode

The function is going to parse the body of the payload, searching for

  • workspace_id
  • workspace_name
  • organization_name
  • run_id

With that information, the function retrieves from Terraform Cloud the last tfstate of the workspace and will save it to S3.

To get all this working, this terraform module terraform-cloud-backup
will create and configure the resources needed.

module "example" {
  source = ""
  bucket_builds  = var.bucket_builds
  bucket_name    = var.bucket_name
  lambda_s3_key  = var.lambda_s3_key
  tf_token       = var.tf_token
  workspaces_ids = var.workspaces_ids
Enter fullscreen mode Exit fullscreen mode

To get this up and running, we need a few stuff.

  • Terraform Cloud Token.
  • The build of the lambda in S3.
  • The name of the S3 bucket where to save the tfstates.
  • The list of the workspaces IDs to backup.

The variables should look like this:
Terraform Cloud Variables

Once that the module has been applied. There is a new notification in Terraform Cloud
Terraform Cloud Notification
And in S3 your tfstates are going to be saved as
tfstate in S3

This is all for today. In the following days, I will publish the Go-Tool to save all the tfstates on Terraform Cloud.

Discussion (0)