Cloudflare Workers Data Platform Integration -- Overview

Executive Summary

Company Manager is a multi-tenant SaaS platform (87 packages, 11 apps, 206+ services) running on Next.js 16 + TRPC + Prisma 7 + Neon PostgreSQL. The platform \1 (chat, rate-limiting, gaming, bookings) with Durable Objects and KV. This document set maps every relevant Cloudflare data platform product to concrete integration opportunities.

Current Architecture Snapshot

Existing Cloudflare Workers

chat-worker        -> https://chat-worker.wd29.workers.dev
rate-limit-worker  -> https://rate-limit-worker.wd29.workers.dev
gaming-worker      -> https://gaming-worker.wd29.workers.dev
bookings-worker    -> https://bookings-worker.wd29.workers.dev

These use Durable Objects, KV, and D1 already. The integration path is well-established.

Document Index

Priority Legend

• **P0** -- Immediate value, low risk, extend existing patterns
• **P1** -- High value, moderate effort, new capabilities
• **P2** -- Strategic value, requires new architecture patterns
• **P3** -- Long-term vision, experimental

Cost Model (Workers Paid Plan -- $5/mo base)

\1: Hyperdrive is free. R2 egress is free. This makes the migration from AWS S3 and direct Neon connections a pure cost win.

Multi-Tenant Considerations

Every integration must respect the existing tenant isolation model:

// TRPC context -- source of truth for tenant scope
ctx.tenantId  // UUID -- REQUIRED in every query
ctx.siteId    // UUID -- optional but recommended
ctx.userId    // UUID -- for user-scoped operations

For Workers:

• **KV**: Use key prefixes `{tenantId}:{key}` or separate namespaces
• **R2**: Use bucket-per-tenant or key prefixes
• **D1**: Use per-tenant databases (D1 supports 50K DBs/account)
• **Vectorize**: Use namespace-per-tenant (50K namespaces/index)
• **Durable Objects**: Include tenantId in DO name/ID
• **Queues**: Single queue with tenant routing, or queue-per-tenant
• **Analytics Engine**: Use index field for tenant segmentation

Integration Architecture

                          Cloudflare Edge
                    ┌─────────────────────────┐
                    │  Workers (compute)       │
                    │  ├─ KV (cache)           │
                    │  ├─ R2 (files)           │
                    │  ├─ D1 (edge SQL)        │
                    │  ├─ DO (state)           │
                    │  ├─ Vectorize (search)   │
                    │  ├─ Workers AI (models)  │
                    │  ├─ Queues (async)        │
                    │  ├─ Analytics Engine      │
                    │  └─ Workflows (durable)   │
                    └──────────┬──────────────┘
                               │ Hyperdrive
                               │ (pooled + cached)
                    ┌──────────▼──────────────┐
                    │  Neon PostgreSQL          │
                    │  (primary database)       │
                    └─────────────────────────┘

    Next.js App (apps/app)
    ├─ TRPC Routers → Workers (via fetch)
    ├─ Service Registry → getService() unchanged
    └─ Prisma → Neon (via Hyperdrive when from Workers)

Risk Assessment

Next Steps

1. Read each design doc for the specific integration you want to implement 2. Follow the \1 for phased rollout 3. Start with P0 items (Hyperdrive, R2, KV) for immediate wins