Migrating existing AWS accounts to the new set up

Migrating existing AWS accounts to the new set up

If you have existing AWS accounts in your Netskope tenant that use the old set up where DLP and malware scans are performed using CloudTrail events, then you can migrate them to the new set up which uses CloudWatch events to efficiently scan your resources for DLP violations and malware.

You can use one of the following methods to migrate your accounts.

Bulk migrate using

You can run the script, migration_script.py to migrate several accounts at the same time. The script requires you to create a CSV file with the following information,

<instance_name>,<api_token>,<tenant-URL>,<aws_access_key_id>,<aws_secret_key>,<aws_region>
<instance_name>,<api_token>,<tenant-URL>,<aws_access_key_id>,<aws_secret_key>,<aws_region>

where,

<instance_name> is the name of the AWS account instance in the Netskope tenant.

<api_token> is the REST API token from the Netskope tenant under Settings > Tools > REST API.

<tenant_url> is the Netskope tenant URL.

<aws_access_key_id> and <aws_secret_key> are the keys from the AWS account.

<aws_region> is the AWS region in which you want the script to upload the migration CFT, aws-instance-setup.yml. For information about the steps performed by the script, see About migration_script.py.

About

This script, migration_script.py performs the following steps.

  1. Reads the CSV file and verifies that the instance for the AWS account exists in the specified Netskope tenant.
  2. Uses the API token to call the REST API and download the migration CFT, aws-instance-setup.yml. The template is customized based on the Netskope Public Cloud services enabled for the account.
  3. Uses the provided AWS access key and secret key to connect to the AWS account and creates a stack called netskopeCAR in the specified AWS region.
  4. Uploads aws-instance-setup.yml to netskopeCAR stack. When the stack creation is complete, the AWS accounts have been migrated to the new set up.

For information on aws-instance-setup.yml and the background process, see “What happens in the process?” sections in the following topics.

Run

To run migration_script.py,

  1. Create a file called migration_script.py and copy the following script into the file.
    import argparse
    import requests
    import time
    import csv
    import json
    import boto3
    import random
    from botocore.exceptions import ClientError
    import concurrent.futures
    from datetime import datetime
    
    
    STACK_UP_TO_DATE = 'No updates are to be performed'
    STACK_NOT_EXIST = 'does not exist'
    MAX_RETRIES = 8
    MIN_STACK_WAIT = 1
    MAX_STACK_WAIT = 10
    CLOUDFORMATION_SERVICE = 'cloudformation'
    STACK_NAME = 'netskopeCAR'
    STACK_SUCC_STATES = ['CREATE_COMPLETE','UPDATE_COMPLETE']
    IN_PROGRESS = 'IN_PROGRESS'
    ROLLED_BACK = 'ROLLBACK_COMPLETE'
    FAILED = 'FAILED'
    STACKS = 'Stacks'
    STACK_STATUS = 'StackStatus'
    STACK_ID = 'StackId'
    STACK_STATUS_REASON = 'StackStatusReason'
    
    APP = "aws"
    PATH = "/api/v1/public_cloud/account"
    
    
    def get_instance(hostname, app, instance_name, token):
        """
        Get instance details
        :param hostname: tenant host
        :param app: app name 'aws'
        :param instance_name: name of the instance
        :param token: REST API token
        :return: instance details or Exception
        """
        print("Getting info for {}".format(instance_name))
        params = {
            "app": app,
            "instance_name": instance_name,
            "token": token
        }
        headers = { "Content-Type" : "application/json", "Accept" : "application/json" }
        get_response = requests.get("https://{}/api/v1/introspection_instance".format(hostname), headers=headers, params=params, verify=False)
        response_data = {}
        json_output = json.loads(get_response.text)
        if json_output.get("status") == "error":
            raise Exception('Errors: {}'.format(json_output.get("errors")))
        response_data["use_for"] = json_output["data"]["instance"]["use_for"]
        response_data["admin_email"] = json_output["data"]["instance"]["admin_email"]
        if "securityscan" in response_data["use_for"]:
            response_data["securityscan_interval"] =  json_output["data"]["instance"]["securityscan_interval"]
        print("Returning info for name {}: {}".format(instance_name, response_data))
        return response_data
    
    def download_cft(hostname, token, app, instance_name, admin_email, usefor, securityscan_interval):
        """
        call the download of the migration CFT for the instance
        :param hostname: tenant host
        :param token: REST API token
        :param app: app name 'aws'
        :param instance_name: instance name
        :param admin_email: admin email for the instance
        :param usefor: services used by the instance
        :param securityscan_interval: security scan interval
        :return: CFT content or Exception
        """
        print("Downloading Migration Template for {}".format(instance_name))
        qparams = {"token":token, "op":"download"}
        headers = { "Content-Type" : "application/json" }
        data = {
            "app": app,
            "type": "cft",
            "use_for": usefor,
            "mode" : "migrate",
            "admin_email" : admin_email,
            "instance_name": instance_name,
            "securityscan_interval": securityscan_interval
        }
        res = requests.post("https://{}{}".format(hostname, PATH), headers=headers, params=qparams, data=json.dumps(data), verify=False)
        response = res.text
        if response.startswith("{"):
            json_output = json.loads(response)
            if json_output.get("status") == "error":
                raise Exception('For instance "{}" errors received: {}'.format(instance_name, json_output.get("errors")))
        return response
    
    def create_resource(instance_name, region, access_key, secret_key, template):
        """
        Call stack creation and wait the stack creation to be completed
        :param instance_name: name of the instance
        :param region: region in which the stack will be created
        :param access_key: access key id for target aws account
        :param secret_key: secret key for the target aws account
        :param template: CFT template
        :return: Exception if encountered
        """
        try:
            print("creating stacks....")
            client = boto3.client(CLOUDFORMATION_SERVICE,
                                        aws_access_key_id=access_key,
                                        aws_secret_access_key=secret_key,
                                        region_name=region)
            if not stack_exists(client, STACK_NAME, region):
                print("stack does not exist, creating stack")
                response = client.create_stack(
                    StackName=STACK_NAME,
                    TemplateBody=template,
                    Capabilities=['CAPABILITY_NAMED_IAM']
                )
                now = datetime.now()
                start_time = datetime.timestamp(now)
            else:
                raise Exception("Stack already present for instance: {}".format(instance_name))
            print("Waiting Stack creation to be completed for {}".format(instance_name))
            while True:
                time.sleep(random.randint(MIN_STACK_WAIT, MAX_STACK_WAIT))
                stacks = client.describe_stacks(StackName=response[STACK_ID])
                stack_status = stacks[STACKS][0][STACK_STATUS]
                if stack_status in STACK_SUCC_STATES:
                    print("Stack creation for {} was successful".format(instance_name))
                    break
                if IN_PROGRESS not in stack_status:
                    if stack_status == ROLLED_BACK or FAILED in stack_status:
                        reason = stacks[STACKS][0].get(STACK_STATUS_REASON)
                        if reason:
                            raise Exception("Exception received while creating stack {} for instance : {} , {}".format(STACK_NAME, instance_name, reason))
                        else:
                            raise
                current_time = datetime.timestamp(datetime.now())
                if start_time + 600 < current_time:
                    raise Exception("Stack creation for {} took longer than expected, please check the AWS console".format(instance_name))
        except ClientError as exc:
            if STACK_UP_TO_DATE not in str(exc):
                raise
            else:
                print("{} Stack {} already up to date and no updates are to be performed".format(instance_name, STACK_NAME))
        except Exception as exc:
            raise
    
    def stack_exists(client, name, region):
        """
        Check if the stack with the provided name is already present
        :param client: client
        :param name: name of the stack
        :param region: region in which the stack is to be deployed
        :return: true if stack exists
        """
        try:
            while True:
                stacks = client.describe_stacks(StackName=name)
                stack_status = stacks[STACKS][0][STACK_STATUS]
                if 'IN_PROGRESS' in stack_status:
                    time.sleep(random.randint(MIN_STACK_WAIT, MAX_STACK_WAIT))
                elif ROLLED_BACK == stack_status or FAILED in stack_status:
                    print("Stack in {} state. Stack cannot be recovered, triggering
                         deleting stack in region : {}".format(stack_status, region))
                    client.delete_stack(StackName=STACK_NAME)
                    time.sleep(random.randint(MIN_STACK_WAIT, MAX_STACK_WAIT))
                else:
                    return True
        except ClientError as exc:
            if STACK_NOT_EXIST in str(exc):
                return False
            else:
                raise
    
    def start_migration(row):
        """
        Caller method for gathering the instance details
        call methods for
        Getting instance details
        Download migration CFT
        call creation of stack
        :param row: single row containing the instance details
        :return:
        """
        details = {
                "instance_name": row[0],
                "apitoken": row[1],
                "tenant_url": row[2],
                "access_key_id" : row[3],
                "secret_key": row[4],
                "region": row[5]
        }
        hostname = details["tenant_url"]
        instance_name = details["instance_name"]
        apitoken = details["apitoken"]
        region = details["region"]
        access_key_id = details["access_key_id"]
        secret_key = details["secret_key"]
        try:
            response_data = get_instance(hostname, APP, instance_name, apitoken)
            use_for = response_data["use_for"]
            admin_email = response_data["admin_email"]
            securityscan_interval = response_data.get("securityscan_interval")
            print("{}, {}".format(use_for, admin_email))
            cft_yaml_body = download_cft(hostname, apitoken, APP, instance_name,
                                         admin_email, use_for, securityscan_interval)
            create_resource(instance_name, region, access_key_id, secret_key, cft_yaml_body)
            print('Migration triggered successfully for : {}'.format(instance_name))
        except Exception as exc:
            return "For {} Exception received : {}".format(instance_name, str(exc))
    
    if __name__ == "__main__":
        """
        Customer will create and provide location of the csv file from which each instance
        migration will be triggered. Migration contains the following parts
        1-> get instance details for the instance
        2-> download migration CFT for the instance
        3-> trigger create of stack with the downloaded CFT
        4-> Wait for the stack creation to be completed
        5-> post success or failure of the migration
        """
        PARSER = argparse.ArgumentParser()
        PARSER.add_argument("--file", help="csv file", required=True)
        ARGS = PARSER.parse_args()
        file_path = ARGS.file
        results = []
        if str(file_path).endswith(".csv"):
            with open(file_path) as f:
                reader = csv.reader(f)
                with concurrent.futures.ThreadPoolExecutor(max_workers=200) as executor:
                    futures = [executor.submit(start_migration, row) for row in reader]
                    for idx, future in enumerate(concurrent.futures.as_completed(futures)):
                        try:
                            res = future.result(timeout=600)
                            results.append(res)
                        except concurrent.futures._base.TimeoutError as exc:
                            print("timeout_error {} {}".format(str(exc), future))
            print("FAILED RESULTS : {}".format(results))
        else:
            print("Expected a csv file and file path: {} is not a csv file".format(file_path))
    
  2. Open a CLI and at the prompt run,
    python migration_script.py --file <path-to-csv>

Manually migrate each account using

You can migrate one account at a time by downloading aws-instance-setup.yml and uploading it to a new CloudFormation stack in the AWS account.

To download aws-instance-setup.yml call the following REST API.

https://<tenant-URL>/api/v1/public_cloud/account?token=<token>&op=download

For more information on REST API endpoints see, Public Cloud API Endpoints for REST API v1.

If you want to download aws-instance-setup.yml through the Netskope tenant,

  1. Go to Settings > Configure App Access > Classic > IaaS page, click on the AWS account to view the edit screen.
  2. Click Start Migration and download the CFT, aws-instance-setup.yml from the migrate account screen.

After you’ve downloaded aws-instance-setup.yml , you must create a new CloudFormation stack in the AWS account and upload the template. For detailed instructions on creating a stack and uploading the template, see the following topics.

Share this Doc

Migrating existing AWS accounts to the new set up

Or copy link

In this topic ...