Optimizely DXP MCP Server

/** * Deployment Workflow Prompts * Provides intelligent prompting for deployment operations and monitoring * Part of Jaxon Digital Optimizely DXP MCP Server * * This module provides MCP prompts to guide AI assistants through proper * deployment workflows, especially continuous monitoring patterns. */ import { PromptDefinition, PromptMessage } from './autonomous-deployment-prompts'; /** * Deployment workflow stage */ export type WorkflowStage = 'start' | 'monitoring' | 'verification' | 'completed' | 'error'; /** * Workflow prompt arguments */ export interface WorkflowPromptArgs { stage?: WorkflowStage; deploymentId?: string; sourceEnvironment?: string; targetEnvironment?: string; } class DeploymentWorkflowPrompts { /** * Get prompt definitions for deployment workflows */ static getPromptDefinitions(): PromptDefinition[] { return [ { name: 'deployment-workflow', description: 'Guide AI through deployment and monitoring workflow', arguments: [ { name: 'stage', description: 'Current stage: start, monitoring, verification, completed, error', required: false }, { name: 'deploymentId', description: 'Deployment ID for context-specific guidance', required: false }, { name: 'sourceEnvironment', description: 'Source environment for deployment', required: false }, { name: 'targetEnvironment', description: 'Target environment for deployment', required: false } ] } ]; } /** * Get prompt messages for a specific workflow stage */ static getPromptMessages(promptName: string, args: WorkflowPromptArgs = {}): PromptMessage[] { if (promptName !== 'deployment-workflow') { throw new Error(`Unknown prompt: ${promptName}`); } const { stage = 'start', deploymentId, sourceEnvironment = 'Preproduction', targetEnvironment = 'Production' } = args; switch (stage) { case 'start': return this.getStartPrompt(sourceEnvironment, targetEnvironment); case 'monitoring': return this.getMonitoringPrompt(deploymentId, targetEnvironment); case 'verification': return this.getVerificationPrompt(deploymentId, targetEnvironment); case 'completed': return this.getCompletedPrompt(deploymentId); case 'error': return this.getErrorPrompt(deploymentId); default: return this.getStartPrompt(sourceEnvironment, targetEnvironment); } } /** * Start stage - after deployment initiated */ static getStartPrompt(sourceEnvironment: string, targetEnvironment: string): PromptMessage[] { return [ { role: 'user', content: { type: 'text', text: `I need to deploy from ${sourceEnvironment} to ${targetEnvironment}` } }, { role: 'assistant', content: { type: 'text', text: `I'll start the deployment from ${sourceEnvironment} to ${targetEnvironment}. Let me initiate this for you. [After starting deployment] ✅ **Deployment Started Successfully!** The deployment is now in progress. I have several monitoring options for you: ## 📊 **Monitoring Options:** ### Option 1: Continuous Monitoring (Recommended) I'll use the \`monitor_deployment\` tool to automatically check the progress every 30 seconds and notify you of important status changes. ### Option 2: Manual Status Checks You can check the status anytime by asking me, and I'll use \`get_deployment_status\`. ### Option 3: One-time Check I'll check once now and you can ask me to check again whenever you'd like. **Which monitoring approach would you prefer?** 💡 **Note:** I'll use the proper MCP monitoring tools instead of bash loops - this is more efficient and provides better progress tracking.` } } ]; } /** * Monitoring stage - during deployment progress */ static getMonitoringPrompt(deploymentId: string | undefined, targetEnvironment: string): PromptMessage[] { return [ { role: 'user', content: { type: 'text', text: 'Yes, please monitor continuously' } }, { role: 'assistant', content: { type: 'text', text: `Perfect! I'll set up continuous monitoring for your deployment using the \`monitor_deployment\` tool. ## 🔄 **Continuous Monitoring Active** **Deployment ID:** ${deploymentId || '[deployment-id]'} **Target:** ${targetEnvironment} **Check Interval:** Every 30 seconds **Max Duration:** 30 minutes I'm now monitoring your deployment in the background. Here's what I'll do: 1. **Check progress** every 30 seconds automatically 2. **Alert you** when the deployment reaches key milestones: - 25% complete - 50% complete - 75% complete - Verification state (needs your review) - Completed 3. **Handle issues** - If any errors occur, I'll notify you immediately **Current Status:** Checking now... [Use \`monitor_deployment\` with deploymentId] You can: - Ask "what's the status?" anytime for an immediate update - Say "stop monitoring" if you want to handle it manually - The monitoring will automatically stop after 30 minutes or when deployment completes I'll keep you updated on the progress! 🚀` } } ]; } /** * Verification stage - when deployment needs verification */ static getVerificationPrompt(deploymentId: string | undefined, targetEnvironment: string): PromptMessage[] { return [ { role: 'system', content: { type: 'text', text: 'The deployment has reached verification state' } }, { role: 'assistant', content: { type: 'text', text: `## ✅ **Deployment Ready for Verification!** Your deployment to ${targetEnvironment} has completed the deployment phase and is now ready for your verification. **Status:** Awaiting Verification (100% deployed) **Deployment ID:** ${deploymentId || '[deployment-id]'} ### 🔍 **Verification Steps:** 1. **Review the changes** at the verification URLs 2. **Test critical functionality** to ensure everything works 3. **Check for any visual or functional issues** ### 📝 **Next Actions:** Once you've verified everything looks good, you have two options: - **Complete the deployment:** Say "complete deployment" or "looks good, complete it" - **Rollback if issues found:** Say "rollback" or "reset deployment" **Note:** The deployment will remain in verification state until you explicitly complete or reset it. There's no automatic timeout. Have you had a chance to review the verification URLs? Let me know when you're ready to proceed!` } } ]; } /** * Completed stage - deployment finished successfully */ static getCompletedPrompt(deploymentId: string | undefined): PromptMessage[] { return [ { role: 'user', content: { type: 'text', text: 'Complete the deployment' } }, { role: 'assistant', content: { type: 'text', text: `## 🎉 **Deployment Completed Successfully!** Your deployment has been finalized and is now live in production. **Deployment ID:** ${deploymentId || '[deployment-id]'} **Status:** Completed **Monitoring:** Ended ### ✅ **What's Done:** - Code successfully deployed to production - Verification completed - Changes are now live - Monitoring has been stopped ### 📊 **Post-Deployment Options:** - View deployment history: \`list_deployments\` - Check environment status: \`status\` - Start a new deployment: \`start_deployment\` Is there anything else you'd like me to help you with?` } } ]; } /** * Error stage - deployment encountered issues */ static getErrorPrompt(deploymentId: string | undefined): PromptMessage[] { return [ { role: 'system', content: { type: 'text', text: 'Deployment encountered an error' } }, { role: 'assistant', content: { type: 'text', text: `## ⚠️ **Deployment Issue Detected** The deployment has encountered an issue that needs your attention. **Deployment ID:** ${deploymentId || '[deployment-id]'} ### 🔧 **Recommended Actions:** 1. **Check detailed status:** Let me get the full deployment status to understand the issue - I'll use \`get_deployment_status\` to get complete details 2. **Review logs:** Check deployment logs for specific error messages 3. **Common resolutions:** - **Timeout issues:** May need to retry the deployment - **Permission errors:** Verify environment access - **Package issues:** Check if the deployment package is valid 4. **Recovery options:** - **Reset deployment:** Use \`reset_deployment\` to rollback - **Retry:** Start a new deployment after addressing the issue Would you like me to get the detailed status to diagnose the issue?` } } ]; } } export default DeploymentWorkflowPrompts;