🏗️ Architecture: How should we classify data across our 4 MongoDB clusters an...

Date: 2026-02-11
Time: 22:50:56 UTC
Session: gspace1-106.11
Agent: architecture
Confidence: 92%
Status: ✅ Documented

📋 Decision Context

Question:

How should we classify data across our 4 MongoDB clusters and 9+ databases for consistent placement of new collections?

Task: Database architecture audit - data classification taxonomy

🎯 Decision Made

Chosen:

Classification by data ownership: User Data → gograb-api, AI System Data → ai-, Vitto Brain → vitto-mind, Tenant Data → space_, Docs → docs-database

Options Considered:

1. Ad-hoc placement based on which backend needs it
2. Classification by data ownership: User Data → gograb-api, AI System Data → ai-, Vitto Brain → vitto-mind, Tenant Data → space_, Docs → docs-database

💭 Reasoning

Data ownership classification provides clear rules for where new collections go: (1) User-owned data (deletable on GDPR request) always goes to gograb-api, (2) AI system intelligence (agent orchestration, decisions, patterns) goes to ai-agents, (3) Vector embeddings go to ai-memory (separated for performance — vector indexes are memory-intensive), (4) MCP/learning/super-intelligence goes to ai-mcp-dev, (5) Vitto's autonomous operational state goes to vitto-mind on its own cluster, (6) Per-tenant content goes to space_* databases, (7) Documentation goes to docs-database on its own cluster, (8) Archived data goes to ai-archive. This taxonomy is documented in docs/english/GoGrab-Space/architecture/database-architecture.md as the source of truth.

📁 Files Involved

• docs/english/GoGrab-Space/architecture/database-architecture.md
• libs/libs-api/backend-core/src/lib/config/database-names.config.ts

🔗 Related

• Decision ID: dev_1770850256813_856seqq
• Session: 106
• Workspace: gspace1

Auto-generated by MCP Smart Consolidation