neptune-quickstart

Purpose

Playing around with Neptune is scary in your personal account, while cheap enough for a business even a small db.t3.medium instance with a ml.t2.medium workbench left running for a month can cost you 70 USD if you forget to turn it off.

The CloudFormation stacks here deploys a Cloud9 instance, a Sagemaker instance, and a Neptune cluster all configured to turn themselves off automatically.

Usage

• - eu-north-1
• - ap-south-1
• - eu-west-3
• - eu-west-2
• - eu-west-1
• - ap-northeast-3
• - ap-northeast-2
• - ap-northeast-1
• - sa-east-1
• - ca-central-1
• - ap-southeast-1
• - ap-southeast-2
• - eu-central-1
• - us-east-1
• - us-east-2
• - us-west-1
• - us-west-2

What's deployed

Cloud9

Cloud9 is an IDE in AWS which is launched in this configuration inside a VPC to allow connectivity to the Neptune instance.

Cloud9 is the easiest to configure to turn itself off, as it turns itself off by default after 30 minutes of inactivity.

Cloud9 requires a service linked role, and an EC2 Instance role with the correct name. As these may already exist in the account, they are created by the helper lambda if they do not exist already.

Sagemaker Notebook

The Sagemaker Notebook in this instance uses the AWS auto-stop-idle script to turn itself off after 30 minutes of inactivity.

Neptune Cluster

Neptune itself can be stopped for a maximum of 7 days before turning itself on again and provides no native way to automate the process of stopping the instance.

This CloudFormation stack deploys a Lambda function which runs every 3 hours (in lambda free tier) and turns off the Neptune cluster if it is currently turned on.

Credits

Adapted from:

• https://github.com/aws-samples/amazon-neptune-samples/tree/master/neptune-workbench-cloudformation
• https://github.com/aws-samples/amazon-neptune-samples/tree/master/neptune-sagemaker/cloudformation-templates/neptune-sagemaker
• https://leehanchung.github.io/2021-06-14-sagemaker-auto-terminate/

Troubleshooting

My CloudFormation stack won't delete!

If you get an error like:

Db cluster neptunedbcluster-wv5radkmww98 is in stopped state (Service: AmazonNeptune; Status Code: 400; Error Code: InvalidDBClusterStateFault; Request ID: 7da76ecf-8752-45db-aa39-935e74beccbb; Proxy: null)

Your CFN Stack won't delete because the Neptune cluster is stopped.
If you start the cluster again (from the AWS console) and retry the stack deletion once the status is Available then it should delete successfully.