iss-anomaly-detection-infra

Overview

Terraform infrastructure to deploy a cloud-based anomaly detection system for monitoring International Space Station telemetry data.

This repository contains Terraform configuration files to deploy an anomaly detection system in AWS that monitors the ISS Loop A cooling system variables (pressure, temperature, flowrate) using a Random Cut Forest algorithm.

The system includes:

• Airflow for pipeline orchestration
• Grafana monitoring dashboard
• Email notifications via Gmail for detected anomalies

Estimated monthly cost: ~150€

Prerequisites

• Terraform 1.10.5 or latest compatible version
• AWS user with administrator permissions
• AWS credentials configured locally

Setup Instructions

1. Configure AWS Credentials

export AWS_ACCESS_KEY_ID=*your key*
export AWS_SECRET_ACCESS_KEY=*your key*

2. Create Backend Resources

• Create an S3 bucket for Terraform state storage
• Create a DynamoDB table for Terraform state locking

3. Configure Backend

Set backend configuration in backend.config:

bucket         = "your unique bucket name"
key            = "terraform/state.tfstate"
region         = "your aws region"
encrypt        = true
dynamodb_table = "your table name"

4. Deploy Infrastructure

cd ./infra
terraform init --backend-config=backend.config
terraform apply

Required Variables

Variable	Description	Type	Sensitive
grafana_admin_password	The admin password for Grafana	string	Yes
airflow_user_password	The user password for Airflow	string	Yes
airflow_user_name	The username for Airflow	string	No
notification_sender_app_password	The app password for Gmail	string	Yes
notification_sender_email	The email address for sending notifications	string	No
notification_receiver_email	The email address for receiving notifications	string	No
aws_account_id	The AWS account ID	string	No
aws_region	The AWS region	string	No
terraform_s3_bucket	The S3 bucket for Terraform state	string	No
terraform_lock_table	The DynamoDB table for Terraform state locking	string	No

Outputs

After deployment, you'll receive the following URLs:

airflow_admin_panel_url = "URL to access the Airflow web interface for pipeline management"
grafana_admin_panel_url = "URL to access the Grafana admin panel for dashboard configuration"
grafana_dashboard_public_url = "Public URL to view the anomaly detection monitoring dashboard"

Teardown

To remove all deployed infrastructure and avoid ongoing costs:

cd ./infra
terraform destroy

Warning: This will permanently delete all resources including:

• EC2 instances and associated storage
• S3 buckets and their contents
• CloudWatch logs and metrics
• All monitoring data and configurations

Make sure to backup any important data before running the destroy command.