MCP DadosBR

# 🚀 Deployment Instructions ## Step 1: Set up GitHub Secret (CI/CD) The CI/CD workflow needs a Cloudflare API token to deploy. ### Get Cloudflare API Token 1. Go to https://dash.cloudflare.com/profile/api-tokens 2. Click "Create Token" 3. Use "Edit Cloudflare Workers" template 4. Select your account 5. Click "Continue to summary" 6. Click "Create Token" 7. **Copy the token** (you won't see it again) ### Add to GitHub Secrets 1. Go to: https://github.com/cristianoaredes/mcp-dadosbr/settings/secrets/actions 2. Click "New repository secret" 3. Name: `CLOUDFLARE_API_TOKEN` 4. Value: Paste your Cloudflare API token 5. Click "Add secret" ## Step 2: Set up Cloudflare Secret (Worker Runtime) The worker needs the Tavily API key at runtime. ### Set TAVILY_API_KEY ```bash # Set the secret wrangler secret put TAVILY_API_KEY # When prompted, paste: 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 ``` Or set for specific environment: ```bash # Staging wrangler secret put TAVILY_API_KEY --env staging # Production wrangler secret put TAVILY_API_KEY --env production ``` ### Verify Secret ```bash # List secrets (won't show values) wrangler secret list # Or for specific environment wrangler secret list --env staging wrangler secret list --env production ``` ## Step 3: Deploy via CI/CD ### Option A: Automatic Deployment (when you push to main) The workflow will automatically deploy to staging when you push changes to: - `lib/workers/**` - `lib/core/**` - `lib/adapters/**` - `wrangler.toml` - etc. ### Option B: Manual Deployment (via GitHub Actions) 1. Go to: https://github.com/cristianoaredes/mcp-dadosbr/actions/workflows/cloudflare-deploy.yml 2. Click "Run workflow" 3. Select environment: `staging` or `production` 4. Click "Run workflow" The workflow will: - ✅ Build the worker - ✅ Deploy to Cloudflare - ✅ Test health endpoint - ✅ Test MCP endpoint - ✅ Show deployment summary ## Step 4: Verify Deployment ### Staging ```bash # Health check curl https://mcp-dadosbr-aredes-staging.cristianocosta.workers.dev/health # List tools curl -X POST https://mcp-dadosbr-aredes-staging.cristianocosta.workers.dev/mcp \ -H "Content-Type: application/json" \ -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' ``` ### Production ```bash # Health check (custom domain) curl https://mcp-dadosbr.aredes.me/health # List tools curl -X POST https://mcp-dadosbr.aredes.me/mcp \ -H "Content-Type: application/json" \ -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' ``` ## Troubleshooting ### Secret not found in worker ```bash # Re-add the secret wrangler secret put TAVILY_API_KEY --env staging ``` ### Deployment fails with 401 - Check that CLOUDFLARE_API_TOKEN is set in GitHub Secrets - Verify the token has Workers permissions ### Health check fails - Wait 30-60 seconds for deployment to propagate - Check Cloudflare dashboard for errors - View logs: `wrangler tail --env staging` ## Next Steps After successful deployment: 1. ✅ Test all endpoints 2. ✅ Monitor logs: `wrangler tail` 3. ✅ Set up custom domain (production) 4. ✅ Configure alerts in Cloudflare Dashboard --- **Need help?** Open an issue: https://github.com/cristianoaredes/mcp-dadosbr/issues