Demo & Installation video: https://drive.google.com/file/d/1FTf0vcxOEKIxX9ALfde1Fp3w8y8eaZVi/view?t=3

Drive link: https://drive.google.com/drive/folders/1OFazb2ErOyh-anSB0HmsZ2kjdIoFETqO

• Self-hosted API health monitoring system built on AWS serverless components.
• Users create monitors (endpoint URL, method, expected status, timeout, check interval, alert email, failure threshold).
• System runs periodic health checks, stores metrics, and sends email alerts when a monitor becomes unhealthy or recovers.
• All infrastructure is defined in Terraform; deployment and teardown are automated with bash scripts.

• Architecture
• Key Design Decisions
• Prerequisites
• Installation Instructions
• Troubleshooting Guide
• Recap
• Credits

Other documentation:

• Terraform – Infrastructure as code, resources, and outputs.
• Lambda functions – API handler, orchestrator, and worker.
• Bash scripts – install.sh and shutdown.sh.

• Frontend: Static site (HTML/CSS/JS) hosted on S3; uses API Gateway base URL for all API calls.
• API: API Gateway REST API (prod stage) with proxy integration to the api-handler Lambda; routes: GET/POST /monitors, DELETE /monitors/{id}.
• Data: DynamoDB – MonitorConfigs (monitor definitions), HealthMetrics (per-check results with state and consecutive failure count).
• Scheduling: EventBridge rule triggers the orchestrator Lambda on a fixed rate (e.g. every 1 minute).
• Queue: SQS queue holds one message per monitor per run; worker Lambda consumes messages.
• Workers: Worker Lambda performs HTTP checks, updates HealthMetrics, and sends SNS alerts (unhealthy when threshold reached, recovery when back to healthy).
• Alerts: SNS topic with email subscription (alert email); SES used to verify the email address. User must confirm SNS subscription in the email.

• Single EventBridge schedule: One rule drives all monitors; orchestrator queries active monitors and fans out to SQS for parallel worker execution.
• Failure and recovery logic: Consecutive failures increment and cap at threshold; consecutive successes decrement to zero. Unhealthy alert when threshold is first reached; recovery alert when failures return to zero.
• Path normalization in API handler: Request path may include stage (e.g. /prod/monitors); handler strips optional first segment so routes work with or without it; DELETE uses last path segment as monitor ID.
• Terraform outputs as source of truth: Install script reads Lambda names, bucket name, API URL from terraform output -json; no hardcoded resource names in scripts.
• Node 18 and aws-sdk: All Lambdas declare aws-sdk (and worker declares axios) in package.json and ship node_modules in the zip; runtime does not bundle aws-sdk.

• AWS account with credentials configured (aws configure).
• Terraform >= 1.0.
• Node.js >= 18 and npm.
• Bash shell.
• zip (for packaging Lambdas).
• Optional: jq (for reliable parsing of Terraform outputs in install script).

• Run the following commands in your terminal:

git clone https://github.com/Abhiram-Rakesh/API-Monitoring-System.git
cd API-Monitoring-System
chmod +x *.sh
./install.sh

• Script will: check prerequisites; prompt for alert email and region if terraform.tfvars is missing; package all three Lambdas; run Terraform init, plan, apply; deploy Lambda code; deploy frontend to S3; print website URL, API URL, and next steps.
• After install: open the website URL; create a monitor; confirm SNS subscription and SES verification for the alert email; wait for the first health checks (interval depends on check_interval_minutes).
• To destroy all resources: run ./shutdown.sh and type yes when prompted.

• Lambda "Cannot find module 'aws-sdk'": Add aws-sdk to the Lambda’s package.json, run npm install in that Lambda directory, rebuild the zip, and redeploy with aws lambda update-function-code. See Lambda functions.
• HealthMetrics not updating: Confirm orchestrator and worker have aws-sdk in package.json and are deployed; check CloudWatch logs for orchestrator (should log "Found N active monitors") and worker (should log "Checking monitor: ..."); ensure EventBridge rule is enabled and SQS queue is attached to worker.
• SNS subscription "Pending confirmation": Confirm the SNS subscription by opening the email from AWS SNS and clicking the confirmation link; SES verification is separate from SNS subscription confirmation.
• API 403 or 404: Use the full API Gateway URL including the stage (e.g. https://xxx.execute-api.region.amazonaws.com/prod); frontend should use the Terraform output api_gateway_url without appending /prod again.
• DELETE monitor not removing: Ensure API handler is redeployed with path normalization and monitor ID from last path segment; check Lambda logs for the received path and extracted ID.
• Website URL wrong format: Terraform output website_url must use format bucket.s3-website.region.amazonaws.com (dot before region); fix in outputs.tf if it used a hyphen.
• Empty Lambda names in install script: Ensure terraform-outputs.json exists after apply; install script uses get_tf_output (jq or python3); install jq or fix JSON parsing if values are empty.

• Monitors are stored in DynamoDB; health checks run on a schedule and results are written to HealthMetrics.
• Alerts are sent via SNS (email) when a monitor hits the failure threshold and when it recovers to zero failures.
• Deleting a monitor from the UI removes it from MonitorConfigs; no further checks or alerts are queued for it.
• Full deployment and teardown are automated via install.sh and shutdown.sh; details are in the linked docs.