Skip to main content
Glama

Solid Multi-Tenant DevOps MCP Server

README.md7.9 kB
# Solid Multi-Tenant DevOps MCP Server AI-first DevOps tooling for managing your Solid SaaS platform with thousands of tenant instances. ## Overview This MCP server provides comprehensive monitoring and management capabilities for your multi-tenant Solid platform through Claude Desktop. Monitor thousands of company instances, track AI agent performance, handle errors, and provision new tenants - all through natural language conversation with Claude. ## Features ### 🏢 **Tenant Management** (Thousands of Customers) - List all tenant companies with pagination - Get detailed tenant information - Search tenants by name, email, or ID - Platform-wide analytics across all tenants - Bulk health checks for multiple tenants ### 🤖 **AI Agent Monitoring** (Per Tenant) - Monitor all 12 AI agents (ADA, Sarah, Jake, Sophia, Marcus, Alex, Jordan, Maya, Riley, Casey, Annie, Devon) - Track inter-agent communication - View agent status and performance metrics - Monitor token usage and costs per tenant ### 💰 **Token Billing & Analytics** - Track AI token usage per tenant - Monitor costs with 2.5x markup - View budget status and alerts - Cost savings analysis (AI vs Human employees) ### 🔍 **Error Monitoring** - List recent errors across all tenants - Filter by tenant, severity, or time range - Get detailed error information with stack traces - Mark errors as resolved with notes ### ⚡ **Performance Monitoring** - System-wide performance metrics - Identify slow API endpoints - Database query analysis - Real-time health checks ### 🐳 **Docker Management** - Monitor PostgreSQL, Redis, and Backend containers - Check container health status - View container logs ### 🚀 **Tenant Provisioning** - Provision new tenant companies - Configure subscription plans (starter, professional, enterprise) - Enable demo mode with sample data - Automated setup workflow ## Installation ### 1. Install Dependencies ```bash cd solid-mcp-server npm install ``` ### 2. Configure Environment The `.env` file is already created with your local endpoints: ```bash BACKEND_URL=http://localhost:8090 FRONTEND_URL=http://localhost:3000 CMS_URL=http://localhost:3001 SUPER_ADMIN_URL=http://localhost:8080 AGENTS_URL=http://localhost:8091 ``` ### 3. Configure Claude Desktop Add this to your Claude Desktop config file: **macOS**: `~/Library/Application Support/Claude/claude_desktop_config.json` ```json { "mcpServers": { "solid-devops": { "command": "node", "args": ["/Users/adamcampbell/Desktop/Solid/solid-mcp-server/src/index.js"], "env": {} } } } ``` ### 4. Restart Claude Desktop Quit Claude Desktop completely and reopen it. ## Available Tools ### Tenant Management - `list_all_tenants` - List all tenant companies (with pagination) - `get_tenant_details` - Get detailed info about a specific tenant - `get_platform_analytics` - Platform-wide analytics - `search_tenants` - Search by name, email, or ID - `bulk_check_tenant_health` - Health check multiple tenants at once ### System Health - `check_all_services_health` - Check Backend, Agents, Docker services - `check_docker_containers` - Check PostgreSQL, Redis, Backend containers ### Tenant-Specific Monitoring - `get_tenant_agents` - All AI agents for a tenant - `get_tenant_agent_messages` - Inter-agent communication - `get_tenant_products` - Products/inventory for a tenant - `get_tenant_token_usage` - AI token usage and billing ### Error Monitoring - `list_recent_errors` - Recent errors (all tenants or specific) - `get_error_details` - Detailed error information - `resolve_error` - Mark error as resolved ### Performance Monitoring - `get_performance_summary` - Overall system performance - `get_slow_endpoints` - Find slow API endpoints ### Tenant Provisioning - `provision_new_tenant` - Create new tenant company ### Database Operations - `execute_raw_query` - Execute raw SQL (READ-ONLY for safety) ## Usage Examples ### Example 1: List All Tenants ``` You: Show me all tenants in the system Claude will call: list_all_tenants(limit=100, offset=0) ``` ### Example 2: Check Tenant Health ``` You: Check the health of company ID 1, 2, and 3 Claude will call: bulk_check_tenant_health(company_ids=[1, 2, 3]) ``` ### Example 3: Monitor AI Agents ``` You: Show me all AI agents for company 5 and their status Claude will call: get_tenant_agents(company_id=5) ``` ### Example 4: View Recent Errors ``` You: Show me critical errors from the last hour Claude will call: list_recent_errors(severity="critical", limit=50) ``` ### Example 5: Provision New Tenant ``` You: Create a new tenant for Acme Corp with email admin@acme.com on the professional plan Claude will call: provision_new_tenant( company_name="Acme Corp", owner_email="admin@acme.com", plan="professional" ) ``` ### Example 6: Monitor Performance ``` You: Find slow API endpoints that take longer than 2 seconds Claude will call: get_slow_endpoints(threshold_ms=2000, limit=20) ``` ## Architecture ### Multi-Tenant Design - Each tenant is a `Company` with a unique `company_id` - All data is scoped per company/tenant - Agents, products, users, orders all belong to a company - Platform admins can monitor all tenants from http://localhost:8080/admin ### Service Endpoints - **Backend**: http://localhost:8090 - Main API server - **Super Admin**: http://localhost:8080 - Platform management dashboard - **Agents**: http://localhost:8091 - AI agent execution server - **Frontend**: http://localhost:3000 - Customer-facing UI - **CMS**: http://localhost:3001 - Public CMS ### Database - **PostgreSQL**: Company data, agents, products, orders - **Redis**: Caching, event bus, performance monitoring ## Workflow Examples ### Morning Health Check ``` 1. Check all services health 2. View platform analytics 3. Check for critical errors overnight 4. Review slow endpoints that need optimization ``` ### Tenant Onboarding ``` 1. Provision new tenant 2. Verify agents are created 3. Check initial setup completed 4. Monitor first customer interactions ``` ### Performance Investigation ``` 1. Get performance summary 2. Identify slow endpoints 3. Check specific tenant experiencing issues 4. Review error logs for that tenant ``` ### Scaling Monitoring ``` 1. List all tenants with pagination 2. Bulk check health of top 50 tenants 3. Review token usage trends 4. Identify tenants needing plan upgrades ``` ## Troubleshooting ### MCP Server Not Appearing in Claude Desktop 1. Check config file location: `~/Library/Application Support/Claude/claude_desktop_config.json` 2. Verify path to index.js is absolute (not relative) 3. Restart Claude Desktop completely 4. Check logs in Claude Desktop Developer Tools ### Connection Errors 1. Ensure all Solid services are running: `./start-all.sh` 2. Check services are accessible: ```bash curl http://localhost:8090/ curl http://localhost:8091/health ``` 3. Verify Docker containers are running: ```bash docker ps ``` ### Tool Execution Errors - Check the specific endpoint exists in your backend - Some tools may return "Not Found" if the endpoint isn't implemented yet - Backend logs: `tail -f logs/backend.log` ## Development ### Add New Tools Edit `src/index.js` and: 1. Add tool definition to `ListToolsRequestSchema` handler 2. Add tool implementation to `CallToolRequestSchema` handler 3. Test with Claude Desktop ### Update Endpoints Edit `.env` file to point to different service URLs (staging, production, etc.) ## System Requirements - Node.js 18+ - Running Solid services (Backend, Agents, Docker) - Claude Desktop ## Support For issues or questions: - Check backend logs: `tail -f logs/backend.log` - Check agent logs: `tail -f logs/agents.log` - View Docker logs: `docker-compose logs -f` --- **Status**: ✅ Ready for AI-first DevOps monitoring of thousands of tenant instances **Version**: 1.0.0

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/Adam-Camp-King/solid-mcp-server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server