📝 Fix: How to handle production issues from Session 100 (MCP spam, ...

Date: 2026-02-09
Time: 22:38:52 UTC
Session: gspace1-102.2
Agent: backend
Confidence: 95%
Status: ✅ Documented

📋 Decision Context

Question:

How to handle production issues from Session 100 (MCP spam, WS reconnect, prompt assets, Mongoose warnings, duplicate indexes)?

Task: Session 102: Vitto production fixes + intelligence wiring

🎯 Decision Made

Chosen:

6 targeted fixes across backend-ai-api and backend-vitto-api: MCP default false, sentinel reconnect, prompt asset glob, suppressReservedKeysWarning, remove redundant indexes

Options Considered:

• No alternatives documented

💭 Reasoning

Each fix touches exactly 1 file, all independent. MCP was spamming on Railway because it defaulted to enabled. WS gave up after 50 retries losing all real-time events. Prompt files not copied to dist meant fallback defaults always used. Mongoose warnings polluted logs. Duplicate indexes waste storage and slow writes.

📁 Files Involved

• apis/backend-ai-api/src/app/workflow/services/mcp-integration.service.ts
• apis/backend-vitto-api/src/app/connectivity/handoff-ws-client.service.ts
• apis/backend-ai-api/project.json
• apis/backend-ai-api/src/app/workflow/schemas/configuration-session.schema.ts
• apis/backend-ai-api/src/app/schemas/ai-infrastructure.schema.ts
• libs/libs-api/session-repository/src/lib/schemas/workspace-state.schema.ts

🔗 Related

• Decision ID: dev_1770676732418_511hb5l
• Session: 102
• Workspace: gspace1

Auto-generated by MCP Smart Consolidation