Cloud Migration Playbook: Step-by-Step Checklist and Rollback Template for Moving Websites to Managed Cloud Hosting
cloud hostingwebsite migrationmanaged cloud hostingrollback planDNSWordPressdeployment

Cloud Migration Playbook: Step-by-Step Checklist and Rollback Template for Moving Websites to Managed Cloud Hosting

NNumberOne Cloud Editorial Team
2026-11-11
8 min read

A practical cloud migration checklist covering DNS, databases, SSL, testing, cutover monitoring, and a reusable rollback runbook.

Moving a website to managed cloud hosting is safer when treated as a controlled release rather than a single server change. This playbook gives you a repeatable migration checklist, DNS and email safeguards, staging tests, database synchronization options, and a rollback template you can reuse before each cutover.

Overview

A cloud migration changes more than where files are stored. The move can affect application configuration, database access, scheduled jobs, caching, SSL certificates, DNS records, email delivery, webhooks, and third-party integrations. The objective is not simply to copy a website; it is to prove that the new environment can serve the same production workload and that you can return to the old environment if a critical problem appears.

This guide assumes the current site remains available while the new managed cloud hosting environment is prepared. Exact console labels and commands vary by provider, operating system, CMS, and deployment method, so confirm syntax and supported versions in your provider documentation before running production commands.

Migration readiness checklist

  • Record the current hosting configuration, PHP or runtime version, CMS version, plugins, themes, extensions, cron jobs, environment variables, and external services.
  • Confirm ownership and access for the domain registrar, DNS provider, current host, new host, database, storage, email platform, CDN, SSL, analytics, and payment or integration accounts.
  • Take a tested backup of website files and the database. Keep at least one copy independent of both hosting environments.
  • Identify a migration owner, reviewer, communication channel, maintenance window, and explicit go or no-go criteria.
  • Confirm that the destination has enough CPU, memory, storage, database capacity, bandwidth, and log retention for the expected workload.
  • Make a list of items that must not be changed during the migration, such as DNS email records, payment settings, or production API credentials.

For architectural decisions, compare the operational trade-offs in VPS vs cloud hosting. If the site is WordPress, check runtime compatibility before copying production data using the WordPress PHP version compatibility checklist.

What to track

Track observable values before, during, and after the move. A simple migration log is more useful than relying on memory during a stressful cutover.

1. Application and infrastructure inventory

Record the source and destination values for runtime versions, web server type, database engine and version, document root, file permissions, upload limits, memory limits, object caching, page caching, persistent storage, scheduled tasks, and outbound email configuration. Note any hard-coded hostnames, IP addresses, file paths, or environment-specific settings. For a WordPress site, include active plugins, themes, media storage, rewrite rules, and the administrator login path.

2. Data freshness and database consistency

Decide how you will handle content or transactions created during the copy. A static brochure site may tolerate a final file and database export. A busy ecommerce or membership site usually needs a synchronization plan so orders, accounts, comments, or form submissions are not lost.

For a short maintenance window, place the application in maintenance mode, create a final database export, import it to the destination, copy changed files, and validate before reopening traffic. For longer migrations, perform an initial import, then use a provider-supported replication or repeatable incremental synchronization method. Do not assume that copying files alone synchronizes a database.

Illustrative commands, to be adapted for your environment, might look like:

rsync -avz --delete ./site/ user@destination:/var/www/site/
mysqldump -u DB_USER -p DB_NAME > site-backup.sql
mysql -u DB_USER -p DB_NAME < site-backup.sql

Test exports and imports on a non-production copy first. Protect credentials, avoid placing passwords in shell history where possible, and confirm whether your database uses options that require a different export process.

3. DNS, SSL, and email records

Document every relevant DNS record, including A, AAAA, CNAME, MX, TXT, SPF, DKIM, DMARC, verification, and service records. Lower the TTL for web records before the cutover, subject to your DNS provider's behavior. A practical schedule is to use the normal TTL well in advance, reduce it to approximately 300 seconds one or two days before the change, keep it low through validation, and restore a normal value after the migration is stable. Lower TTL does not instantly remove cached responses, so allow for existing caches.

Do not change MX records merely because the website is moving. Email may remain with a separate provider. Preserve authentication records and test both inbound and outbound mail. Plan certificate issuance on the destination before changing DNS, and verify that HTTPS redirects, certificate coverage, renewal, and any www or apex hostname behavior are correct. Use the domain, DNS, SSL, and email setup guide as a companion checklist.

4. Staging and production test results

AreaStaging testProduction check after cutover
Pages and assetsCheck representative pages, images, downloads, redirects, and search functionality.Test top landing pages and recently changed content from more than one network.
DatabaseVerify records, queries, forms, logins, and content editing.Confirm new submissions or transactions appear in the destination database.
SecurityCheck HTTPS, headers, access controls, secrets, and administrative routes.Inspect certificate, redirects, logs, and alerts without exposing sensitive data.
PerformanceRun representative requests and a controlled load test.Compare response times, errors, CPU, memory, storage, and database health.
OperationsRun backups, restores, cron jobs, queues, webhooks, and notifications.Confirm monitoring, alert routing, and the on-call owner.

For performance validation, use the repeatable process in cloud hosting load testing and compare resource readings with the resource monitoring checklist.

Cadence and checkpoints

Use the following workflow as a migration runbook. Add an owner and timestamp beside every item.

Preparation: several days before cutover

  1. Freeze nonessential infrastructure changes and record the current site health baseline: page checks, error rate, traffic pattern, response time, disk usage, and database size.
  2. Build the destination environment and deploy the application without directing public traffic to it.
  3. Import a recent backup and configure environment variables, permissions, queues, cron jobs, caching, and integrations.
  4. Issue or prepare the SSL certificate and test the destination through a temporary hostname or local hosts-file entry.
  5. Lower DNS TTL for the relevant web records and verify that email records remain unchanged.
  6. Run staging tests, restore tests, and a controlled load test. Resolve failures before scheduling the cutover.

Cutover window

  1. Announce the start of the window and enable maintenance mode or write protection if data consistency requires it.
  2. Take a final source backup. Complete the final database synchronization and copy files changed since the initial transfer.
  3. Run smoke tests against the destination using its temporary address. Confirm the expected database version and current content.
  4. Change only the planned web DNS records. Record the exact old and new values and the change time.
  5. Monitor DNS resolution from multiple networks while testing HTTPS, redirects, forms, login, checkout, media, APIs, webhooks, and email.
  6. Keep the source environment intact and writable only if your data strategy explicitly supports it. Otherwise, prevent split-brain writes.

After cutover

For the first hours, monitor application errors, access logs, database connections, queue depth, CPU, memory, storage, SSL status, and business-critical transactions. Keep the old environment available for the agreed observation period. Do not delete it immediately after the first successful page load.

For larger sites, consider blue/green or canary delivery. In a blue/green approach, the old environment is one complete production version and the new environment is the other; traffic switches between them. A canary approach sends a small, controlled portion of traffic to the destination first. These patterns require reliable routing, session handling, shared or synchronized data, and clear monitoring, so they are not automatically safer for every site.

How to interpret changes

A successful DNS response does not prove a successful migration. Interpret results by separating routing, application, data, and capacity problems.

  • DNS or certificate symptoms: inconsistent destinations, certificate warnings, redirect loops, or failures on one hostname suggest incorrect records, propagation, certificate coverage, or proxy configuration.
  • Application symptoms: server errors, missing assets, broken logins, or failed forms often indicate incompatible runtime versions, permissions, missing extensions, incorrect environment variables, or rewrite rules.
  • Data symptoms: missing orders, stale content, duplicate submissions, or failed writes indicate an incomplete final sync, split-brain writes, or a database configuration issue. Pause writes and investigate rather than repeatedly copying data.
  • Performance symptoms: rising latency with normal traffic may point to insufficient resources, uncached requests, database queries, storage performance, or a disabled optimization. Compare like-for-like traffic and cache state.
  • Operational symptoms: absent backups, cron jobs, alerts, or logs mean the migration is incomplete even if visitors can browse the site.

Define rollback triggers before cutover. Examples include persistent errors on a critical user journey, incorrect or missing transactional data, inability to issue valid HTTPS, failed database writes, or a security control that cannot be verified. A temporary increase in latency may be investigated without rollback if core journeys work and the trend is improving; a data integrity problem generally deserves a faster response.

Editable rollback template

Migration: [site and environment]
Owner: [name]    Reviewer: [name]
Cutover start: [UTC date/time]    Decision deadline: [UTC date/time]
Source environment: [hostname/IP]
Destination environment: [hostname/IP]
DNS records changed: [record names and old/new values]
Rollback triggers: [critical errors, data, SSL, security, or availability conditions]

Rollback steps
1. Declare rollback in [channel] and record the reason and timestamp.
2. Stop writes or enable maintenance mode to prevent data divergence.
3. Restore planned web DNS records to the source values.
4. Confirm source application, database, SSL, and integrations are healthy.
5. Validate representative pages and critical transactions on the source.
6. Monitor DNS, errors, and traffic until the source is serving normally.
7. Preserve destination logs, database state, and change records for analysis.
8. Reconcile any writes made during the transition before a later retry.

Go-live evidence: [links to tests, logs, backup, and monitoring]
Rollback result: [successful/partial/failed]
Follow-up owner and deadline: [details]

Backups are useful only if restoration is understood and tested. Before migration, review the backup and restore workflow, and after migration confirm that the new provider's backup schedule and retention match your operational needs.

When to revisit

Revisit this playbook before every hosting change, major CMS or plugin upgrade, domain change, architecture change, or launch that alters traffic or data behavior. Even without a planned migration, review the runbook monthly or quarterly: verify access, backup restoration, DNS ownership, SSL renewal, monitoring, cron jobs, resource trends, and the contact responsible for rollback.

Refresh provider-specific console steps, CLI examples, screenshots, migration-tool versions, and CMS compatibility notes whenever a provider interface or supported software version changes. After an incident, compare the validation log with the actual failure, add the missing test, and update the rollback trigger. This turns the document into an operational record rather than a one-time checklist.

Before your next migration, copy the template into your change-management system, assign an owner to each checkpoint, capture baseline measurements, and schedule a short post-cutover review. Keep the old environment and final backup until the observation period is complete and the team has confirmed that data, security, performance, and recovery procedures all work as intended.

Related Topics

#cloud hosting#website migration#managed cloud hosting#rollback plan#DNS#WordPress#deployment
N

NumberOne Cloud Editorial Team

Cloud Hosting Editors

Senior editor and content strategist. Writing about technology, design, and the future of digital media. Follow along for deep dives into the industry's moving parts.