Carbone.io Middleman with Reporting & Cost Tracking

Overview

Carbone.io is a powerful, template-based document generation engine that fits seamlessly into serverless architectures. It’s excellent at producing PDFs on the fly — but it lacks built-in reporting. That limitation became a launch point for this internal project: building a middleman service layer to integrate Carbone with Azure Logic Apps and a PostgreSQL database for rich reporting and cost tracking.

Project Goals

1. Create a serverless, reusable middleware that simplifies Carbone API usage.
2. Enable metadata logging of each PDF generation (team, client, project, etc.).
3. Introduce cost-awareness by tracking usage volume per team.
4. Provide reportable data in Metabase for visibility and accountability.

Architecture Overview

The solution uses three Azure Logic Apps, each performing one critical step in the Carbone document lifecycle:

• create-template: Upload and register the document template.
• generate-document: Merge data into a chosen template to generate a PDF.
• render-document: Retrieve and return the final document.

All three steps interact with a PostgreSQL database, where metadata from each operation is logged in real time.

Middleman Parameters

Every document operation is tagged with:

• team
• client
• project
• product
• source

These parameters allow detailed filtering and cost attribution. For example, setting source to “Postman” excludes test runs from billing.

User Workflow

1. Upload Template

Users call the create-template Logic App with form-data, specifying team, template, and optionally client, project, and more.

Response: A templateId for future use.

2. Generate Document

Users post a JSON payload with:

• Data matching their template variables.
• Metadata (team, client, project, etc.).
• Their templateId.

Response: A renderId used to retrieve the final document.

3. Download Document

Users send a simple JSON to the render-document Logic App with renderId and team. The response is the generated PDF.

Reporting Layer

All logic app executions write logs to PostgreSQL. This data feeds a Metabase dashboard, showing:

• Total runs per team or client.
• Effective document cost calculations.
• Monthly usage trends.

Cost Breakdown Example:

Dashboard Overview

Included in the dashboard are:

• Template creation logs.
• Generation statistics by team, client and month.
• Effective rates accross the organisation.

Challenges & Learnings

• Data Structure Design: Mapping parameters like project and source consistently proved vital for meaningful reports.
• Cost Attribution: Carbone’s volume-based pricing made cost division per client nuanced — this solution handles it transparently.
• Reliability: Azure Logic Apps provided retry policies and failure alerting out of the box.

Outcome

The middleware allows each team to confidently:

• Track document generation activity.
• Attribute costs to the right clients.
• Have transparency around costs.

Tech Stack

Conclusion

This project filled a critical gap between a powerful tool (Carbone.io) and enterprise requirements (reporting, billing, traceability). The middleware ensures visibility, accountability, and cost-effectiveness.