🏗️ Architecture: Do we need to modify the git-based Railway deployment proces...

Date: 2026-01-25
Time: 21:08:07 UTC
Session: editor2-39.40
Agent: backend
Confidence: 95%
Status: ✅ Documented

📋 Decision Context

Question:

Do we need to modify the git-based Railway deployment process for new claude-agent-mcp features?

Task: Not specified

🎯 Decision Made

Chosen:

No changes needed - existing process works as-is

Options Considered:

• No alternatives documented

💭 Reasoning

The git-based deployment (push to deploy/ai-api branch) automatically works because: 1) Railway watches the branch and reads railway.toml, 2) railway.toml points to Dockerfile.railway.backend-ai which we already updated, 3) The Dockerfile builds both backend-ai-api AND claude-agent-mcp, 4) Environment variables have defaults in Dockerfile. Only recommended change is adding apps/claude-agent-mcp/** to Railway watch paths for independent MCP updates.

📁 Files Involved

• docs/english/Deployment/Internal/Railway/github-branches.md
• deployment/Dockerfile.railway.backend-ai
• deployment/railway.backend-ai.toml

🔗 Related

• Decision ID: dev_1769375287002_zksjnvk
• Session: 39
• Workspace: editor2

Auto-generated by MCP Smart Consolidation