deployment-procedures

Deployment Procedures - Production Deployment Principles and Decision Guide

Skill Overview

Deployment Procedures is a skill focused on production environment deployment principles and decision-making mindset. It helps developers master secure deployment workflows, rollback strategies, and validation methods, teaching them to think rather than rote-memorize scripts.

Applicable Scenarios

1. Production Application Releases

When you need to deploy an application to production, this skill provides a complete pre-deployment checklist, execution process, and post-deployment verification methods to ensure the release is safe and controllable.

2. Multi-Platform Deployment Strategy Selection

Whether using hosting platforms like Vercel/Netlify, the traditional VPS+PM2 approach, or containerized deployments with Docker/Kubernetes, this skill helps you choose the appropriate deployment strategy and rollback plan.

3. Emergency Incident Fast Recovery

When a service interruption or serious error occurs in production, this skill provides emergency handling procedures to help you quickly assess the problem, execute a rollback or fix, and minimize impact.

Core Features

1. Platform-Adapted Deployment Processes

Provides targeted deployment methods and rollback strategies for different platforms (Vercel, Railway, Docker, Kubernetes, etc.), avoiding one-size-fits-all processes and ensuring each platform uses best practices.

2. Five-Stage Deployment Workflow

Breaks the deployment process into five stages: Prepare, Backup, Deploy, Verify, Confirm/Rollback. Each stage has clear checklists and execution principles to reduce deployment risk.

3. Zero-Downtime Deployment Strategies

Explains in detail three zero-downtime strategies—rolling updates, blue-green deployments, and canary deployments—covering applicable scenarios and implementation methods to update versions while ensuring service continuity.

Frequently Asked Questions

What preparations are needed before production deployment?

Complete verification in four areas before deployment: code quality (tests passing, code review), build validation (production build succeeds without warnings), environment checks (environment variables and keys configured correctly), and security preparation (backups completed and a rollback plan established). It is recommended to use a pre-deployment checklist to confirm each item and avoid missing critical steps.

How to quickly roll back after a failed deployment?

Rollback speed takes priority over perfect analysis. Rollback methods vary by platform: Vercel/Netlify can redeploy the previous commit; Railway/Render offer one-click rollback in the dashboard; VPS+PM2 requires restoring backups and restarting services; Docker uses the previous image tag; Kubernetes runs kubectl rollout undo. The rollback principle is: restore service first, then analyze the cause; avoid multiple attempts while the system is in a faulty state.

When should production deployments be scheduled?

Best practice is to deploy during mid-morning on workdays (Tuesday to Thursday mornings), avoiding Friday afternoons or weekends. The reason is simple: deployments require at least 15 minutes of active monitoring, and if issues arise the team should be available to respond during working hours. Deploying on Friday can affect the weekend, and weekend deployments lack support personnel. Also recommended are small-batch, high-frequency releases rather than large one-time changes.