The Next Job for AWS Database Migration Services Is Already on AWS
AWS Cloud
AWS Cloud


When most organizations hear "database migration," they picture a move from an on-premises data center to the cloud. That is how AWS migration services are usually sold: assess the source databases, choose a target engine, replicate with AWS Database Migration Service, and cut over. Guides to AWS database migration services focus on replication instances, change data capture, and homogeneous versus heterogeneous moves.
That work still matters. But for companies that moved to Amazon RDS years ago, the most pressing migration now sits inside AWS. Amazon RDS for MySQL 8.0 reached the end of standard support on July 31, 2026. Databases enrolled in RDS Extended Support, the default for many existing instances, began incurring Extended Support charges on August 1. Every one of those databases needs a major version upgrade, and the upgrade to MySQL 8.4 is not trivial.
The next job for AWS database migration services is already on AWS. Migration services should be scoped and costed for existing RDS estates, with the same rigor applied to on-premises moves, because the costs of waiting now appear on the monthly bill.
What Changed on July 31, 2026
RDS for MySQL 8.0 was planned to reach end of standard support on July 31, 2026. RDS Extended Support is a paid offering that provides critical security and bug fixes for up to three years after that date. It is charged per vCPU per hour, varies by Region, and rises in year three. Multi-AZ instances are billed for both primary and standby.
AWS's own examples for a db.r5.xlarge in US East illustrate the scale:
Single-AZ (4 vCPUs): about $292 a month in years one and two, $584 in year three
Multi-AZ (8 vCPUs): about $576 a month, then $1,152
Multi-AZ with a Multi-AZ read replica (16 vCPUs): about $1,168 a month, then $2,336
The RDS Extended Support documentation explains that enrollment status shows in the console, CLI, and API, and that charges for MySQL 8.0 databases enrolled in Extended Support start on August 1, 2026. Enrollment is the default unless it was disabled when the database was created.
For a company running 40 or 50 MySQL databases of various sizes, many with replicas, Extended Support can add tens of thousands of dollars a month. It buys time, but it does not remove the upgrade.
Why AWS Database Migration Services Now Cover Upgrades
Major version upgrades can look simple in the console: choose a new version and apply. In practice, MySQL 8.0 to 8.4 involves changes that can break applications:
Authentication: mysql_native_password is disabled by default, and caching_sha2_password becomes the default. Older drivers and some tools may fail to connect.
Terminology: SOURCE and REPLICA replace MASTER and SLAVE in replication commands, so scripts and automation need updating.
Reserved keywords: new reserved words can break queries that use them as identifiers.
InnoDB defaults: several system variable defaults change, including innodb_adaptive_hash_index, which can affect performance.
Each of these must be found, fixed, and tested across every application that touches the database. That is migration work: inventory, assessment, remediation, testing, and cutover.
Driver and Application Readiness
The authentication change deserves particular attention because it fails loudly and immediately. When mysql_native_password is disabled, any client that cannot negotiate caching_sha2_password stops connecting the moment the database switches over. Older versions of language connectors, some business intelligence tools, legacy ETL software, and embedded clients in vendor applications are common culprits.
The fix is usually straightforward, such as upgrading a driver or changing a connection setting, but only if the client is known. That is why dependency mapping matters. Database logs, connection metadata, and security group rules can reveal which hosts connect to each database, and application owners can confirm the driver versions in use. Vendor applications need special care, since the company may not control their connectors and may need the vendor to confirm compatibility or provide an update.
Testing Performance Before Switchover
Changed InnoDB defaults and optimizer behavior can shift query performance in either direction. Most workloads see little change, but a few critical queries may slow down or choose different plans. Capture a representative workload from production, replay it against a blue/green copy on 8.4, and compare response times for the most important queries. Review custom parameter groups and decide deliberately whether to keep old values or adopt new defaults. Teams that skip this step sometimes discover performance problems during peak hours after switchover, when rollback is hardest.
The Arithmetic of Waiting
Extended Support costs scale with vCPUs and grow in year three. A company with ten Multi-AZ production databases similar to AWS's example would pay roughly $5,760 a month in years one and two, or about $69,000 a year, before counting replicas and non-production instances. In year three, that figure roughly doubles. Against those numbers, a well-run upgrade program that finishes within two or three quarters usually pays for itself, while also restoring the security posture of a version under standard support.
Choosing an Upgrade Strategy
AWS describes three main paths, each with trade-offs:
In-place upgrade: The simplest path, keeping the same endpoint and storage. Downtime ranges from minutes to hours, and the upgrade cannot be canceled once started, although RDS takes a snapshot first.
Blue/green deployment: RDS builds a synchronized copy on the new version, and switchover typically takes under five seconds, with the option to switch back. It requires automated backups and costs more during the deployment window.
Snapshot restore: A manual approach suited to development and test, with a new endpoint and replicas to recreate.
For production databases with tight downtime limits, blue/green deployments are often the right choice. For less critical databases, in-place upgrades during a maintenance window may be enough. The decision should be made database by database.
What AWS Migration Services Should Include for RDS Estates
A well-scoped engagement treats the RDS estate as a migration portfolio:
Inventory: Every RDS and Aurora database, engine version, size, replica configuration, Extended Support enrollment, and standard support end date.
Cost exposure: Current and projected Extended Support charges per database, by month, to prioritize the work.
Dependency mapping: Applications, services, ETL jobs, reporting tools, and integrations that connect to each database, with their drivers and authentication methods.
Compatibility assessment: Checks for deprecated authentication, reserved words, replication commands, and parameter group settings.
Remediation: Driver upgrades, query fixes, script updates, and parameter adjustments.
Testing: Functional and performance testing on the new version, using blue/green or restored copies.
Cutover planning: Strategy per database, maintenance windows, rollback plans, and communication.
Post-upgrade validation: Monitoring for errors and performance changes after switchover.
Migration services in AWS that cover these steps turn a stressful deadline into a managed program.
A Representative Estate
Consider a software company that moved to RDS in 2019 and now runs 32 MySQL 8.0 databases: eight production Multi-AZ instances with read replicas, twelve staging and test instances, and twelve smaller databases for internal tools. All are enrolled in Extended Support by default.
An inventory shows that Extended Support for the production databases alone adds several thousand dollars a month in years one and two, and twice that in year three. Dependency mapping finds three legacy services using an old connector that does not support caching_sha2_password, and two reporting jobs that reference a column named with a newly reserved word.
The company upgrades test and internal databases in place within a month, fixes the connector and queries, then moves production databases one by one using blue/green deployments over the following quarter. Extended Support charges fall each time a database moves to 8.4.
Who Owns the Upgrade
Engine upgrades often fall between teams. Platform engineers manage the RDS instances, application teams own the code that connects to them, data teams run pipelines and reports, and vendors supply packaged software. Each group assumes another will handle the version change. Without a named owner, the work drifts while Extended Support charges accumulate.
Give the program a single owner with authority to schedule changes across teams, and agree a simple service level for each application team: confirm drivers and queries within two weeks of being asked, and join testing for their databases. An outside migration partner can run the program, but internal owners must still approve cutovers and accept the results.
Do Not Forget the Other Engines
MySQL 8.0 is the most visible case in 2026, but the pattern repeats. RDS for PostgreSQL, Aurora, and other engines follow their own release calendars, each with standard support end dates and Extended Support options. AWS cloud migration services should maintain a forward calendar of these dates for the whole estate, so the next deadline is planned well in advance rather than discovered on a bill.
AWS Cloud Data Migration Still Has a Place
On-premises and cross-cloud migrations continue, and AWS Database Migration Service remains central to them. An engine upgrade can also be an opportunity to reconsider architecture: consolidating small databases, moving suitable workloads to Aurora, or changing instance classes. AWS cloud data migration teams that understand both external moves and internal upgrades can combine these decisions into a single plan.
Security and Compliance Considerations
Extended Support provides critical security fixes, so enrolled databases are not left unpatched. Still, auditors and customers increasingly ask about supported versions in security questionnaires. Running on Extended Support for long periods can raise questions in SOC 2, ISO 27001, PCI DSS, and HIPAA reviews. Upgrading to a version under standard support simplifies those conversations. The upgrade itself must preserve encryption, access controls, audit logging, and backup configurations, which should be part of the test plan.
A Practical Starting Point
Start with three steps this month: list every RDS database with its engine version and Extended Support status, calculate the monthly Extended Support cost for each, and identify the applications connected to the most expensive ones. That list sets the order of work. Then test the first upgrade on a non-production copy to uncover compatibility issues early.
The end of standard support for RDS for MySQL 8.0 has turned engine upgrades inside AWS into a priority migration, with Extended Support charges now accruing for databases that remain on 8.0. AWS database migration services should cover existing RDS estates with inventory, cost analysis, compatibility remediation, testing, and cutover planning, alongside traditional on-premises moves. It's best to opt for AWS migration services that scope and deliver both kinds of database migration. Each month on Extended Support is a cost that a planned upgrade can end.
Sign in to respond and applaud.
No responses yet. Be the first.
Software Development
IT Services & Consulting
SaaS / Cloud Software