Waste Radar for AWS

Waste Radar uses AI to detect AWS cost waste and automate remediation in Jira. It turns cloud insights into actionable tasks, and Confluence reports to streamline optimisation.

Comment

Waste Radar for AWS

Inspiration

Cloud cost optimization is often a manual, reactive process. Engineers frequently discover wasteful spending weeks after the fact when the monthly bill arrives. My inspiration was to build a "Cloud Hygiene" bridge directly where engineering teams and management live: the Atlassian ecosystem. I wanted to turn complex cloud insights into immediate, actionable tasks that teams can resolve as part of their daily workflow, rather than an afterthought.

What it does

Waste Radar for AWS is a comprehensive Forge application that proactively identifies cloud cost waste and automates remediation.

  • Zero-Config AWS Simulation: For judges and reviewers, I've integrated a high-fidelity AWS simulation engine. This allows you to see the full power of the app immediately, without needing to configure complex IAM roles or AWS credentials.
  • Scans & Identifies: It analyzes simulated AWS environments to find underutilized resources like unattached EBS volumes, oversized EC2 instances, and inefficient S3 storage tiers.
  • Actionable Remediation: With a single click, you can generate Jira tasks for every finding, complete with risk levels, cost impact, and technical recommendations.
  • AI-Powered Assistant: An integrated Atlassian Rovo Agent, "Waste Radar Assistant," allows you to query cost findings using natural language (e.g., "Show me high-risk EC2 waste").
  • Executive Reporting: It automatically generates professional Confluence reports, summarizing potential monthly savings and linking back to remediation efforts in Jira for leadership visibility.

Quick Start for Judges

  1. Install the app via the provided installation link.
  2. Navigate to any Jira project (e.g., "KAN").
  3. Click the Waste Radar menu item in the left sidebar.
  4. Click Run AWS Scan to populate the dashboard with cost findings.
  5. Use Create Tasks to generate remediation issues in Jira.
  6. Use Generate Report to create an executive summary in Confluence.
  7. Try Rovo: Open the Rovo sidebar and ask "Waste Radar Assistant" to summarize the findings.

Built with

  • Atlassian Forge: The serverless platform used for building and hosting the application.
  • Atlassian Rovo: To create the "Waste Radar Assistant" and custom AI-powered actions.
  • React: For building the interactive custom UI dashboard.
  • Node.js: Powering the backend resolvers and API communication.
  • Vite: For fast frontend development and bundling.
  • Jira Cloud REST API: To programmatically create and manage remediation tasks.
  • Confluence Cloud REST API: To automate the generation of executive cost reports.
  • Atlassian Document Format (ADF): For structured rich-text injection into Jira issues.
  • CSS3: For the custom, high-performance "racing-inspired" styling.

How I built it

The application is built on the Atlassian Forge platform leveraging Custom UI and Rovo:

  • Frontend: A React-based dashboard built with Vite, featuring custom interactive charts and a premium design system tailored for high-stakes engineering environments.
  • Backend: Node.js resolvers using @forge/api to interact with Jira and Confluence Cloud REST APIs.
  • Atlassian Rovo: A specialized Rovo Agent with a custom action (get-waste-summary) that translates infrastructure data into conversational insights.
  • Integration: I mastered the Atlassian Document Format (ADF) for rich Jira descriptions and Confluence Storage Format for dynamic table generation.

Challenges I ran into

  • Cross-Product State Management: Coordinating data across Jira and Confluence while maintaining a seamless user experience in a single Forge app required careful handling of asynchronous API calls and product-specific permissions.
  • ADF Complexity: Designing the rich text formatting for Jira tasks was challenging, as I had to ensure the information was both human-readable and programmatically consistent.
  • Rovo Action Schema: Aligning the Rovo action inputs and outputs with my existing infrastructure data model took several iterations to ensure the AI teammate provided precise, useful advice.

Accomplishments that I'm proud of

  • Unified Workflow: I successfully created a workflow that bridges Infrastructure, Engineering (Jira), and Management (Confluence) in one tool.
  • Impactful AI: Integrating Rovo, the AI assistant, provides a genuine efficiency boost by summarizing complex findings in seconds.
  • Premium UI/UX: I built a dashboard that feels like a professional enterprise tool, featuring a "Williams Racing Edition" visual flair to celebrate peak performance.

What I learned

  • The Power of Forge: Building cross-product apps is a superpower. The ability to push data from one product (Jira) to another (Confluence) seamlessly is incredibly powerful for reporting.
  • AI Context Matters: I learned that the quality of a Rovo Agent is defined by the context and data structure provided to its actions.
  • Developer Agility: Using Forge Custom UI with Vite significantly sped up my development cycle, allowing me to focus on the business logic of cost optimization.

What's next for Waste Radar for AWS

  • Live AWS Integration: Implementing cross-account IAM roles to connect to real AWS Organizations.
  • Direct Remediation: Empowering Rovo to execute cleanup actions (e.g., "Rovo, delete the unattached volumes identified in us-east-1").
  • Historical Savings Tracking: Adding a dashboard to track actual money saved over time as Jira tasks are moved to "Done."

Built With

Share this project:

Updates

Submission history