How to Prepare for AWS Service Discontinuations: A Step-by-Step Guide

Introduction

AWS regularly evolves its service portfolio, sometimes retiring or deprioritizing older services. Recent announcements like the discontinuation of WorkMail, App Runner moving to maintenance mode, and several other services entering sunset phases have sparked concern among users. Understanding how to respond proactively can save you from unexpected disruptions and help you maintain a resilient cloud architecture. This step-by-step guide walks you through the process of handling AWS service discontinuations—from staying informed to migrating your workloads—so you can protect your applications and data.

What You Need

• AWS account with administrative access
• Inventory of services currently in use (AWS Resource Explorer, Config, or manual list)
• AWS Health Dashboard access to track service announcements
• Basic understanding of your workload dependencies (e.g., which services are critical)
• Backup or snapshot of critical data and configurations
• Testing environment to validate replacements before production cutover

Step-by-Step Guide

Step 1: Stay Informed with Official Announcements

The first line of defense is awareness. AWS communicates service discontinuations and maintenance changes through official channels.

• Check the AWS Health Dashboard daily or set up notifications via Amazon EventBridge and AWS Health.
• Subscribe to AWS What’s New and the AWS News Blog for major updates.
• Monitor the AWS Service Health page for live status changes.
• Read your email associated with the AWS account; AWS often sends personalized notifications about affected services.

By doing this, you’ll catch announcements early—often months before the sunset date—giving you time to plan.

Step 2: Assess Impact on Your Architecture

Once you learn about a discontinuation, immediately evaluate which of your resources are affected.

• Use AWS Resource Explorer to run a query for the retiring service (e.g., service:workmail).
• Review AWS Config rules to identify resources that rely on the service.
• Talk to your team about any undocumented usage or shadow IT deployments.
• Document dependency chains: If the retired service feeds data into another system (e.g., Lambda -> WorkMail -> S3), trace the entire flow.
• Quantify the risk: assign a priority based on whether the service is used for production, development, or experiments.

Create a spreadsheet or use AWS Systems Manager to tag all impacted resources. This will be your migration baseline.

Step 3: Evaluate Replacement Options

AWS usually recommends alternative services for retired ones. Research and compare these options against your requirements.

• Review AWS documentation for migration paths. For example, WorkMail users are directed to Amazon WorkMail? (wait – that’s being discontinued; actually, AWS suggests using alternative email services or third-party providers.) Check the official FAQ.
• Consider AWS native alternatives: For App Runner, you could move to AWS Fargate or Amazon ECS.
• Evaluate third-party or open-source solutions if no AWS counterpart fits (e.g., use self-hosted Postfix for email instead of a managed service).
• Test cost and performance of the replacement in a non-production environment. Run a side-by-side comparison.
• Check compatibility: Will the new service integrate with your existing IAM, networking, and monitoring stack? If not, you may need adjustments.

Choose the option that minimizes re‑architecture and operational overhead while meeting compliance and scalability needs.

Step 4: Plan and Execute the Migration

With a clear target in mind, create a detailed migration plan with timelines and rollback procedures.

• Set a deadline based on the AWS sunset date—add a buffer of at least two weeks.
• Back up all data from the retiring service (emails, application logs, etc.) and store them in Amazon S3 or Glacier.
• Automate the migration where possible: use AWS CloudFormation or Terraform to deploy the replacement infrastructure.
• Update DNS, routing, and API endpoints to point to the new service. Test connectivity from all client applications.
• Perform a dry run in a staging environment identical to production. Validate that data flows correctly.
• Cut over in a maintenance window and monitor for errors. Have a rollback plan ready if something fails.

For complex migrations, break the work into phases: move non‑critical workloads first, then migrate production after successful testing.

Step 5: Validate and Clean Up

After migration, verify that everything works and remove dependencies on the retired service to avoid future billing surprises.

• Run end‑to‑end tests on all affected features. Check logs, metrics, and user reports.
• Decommission the old resources only after the sunset date passes and you’re confident the migration is stable. Delete related IAM roles, security groups, and CW alarms.
• Update your architecture documentation to reflect the new service and note any configuration changes.
• Communicate the change to your team and stakeholders. Celebrate the successful migration!
• Set a periodic review cycle (quarterly) to scan for upcoming AWS retirements using the Health Dashboard.

This final step ensures you don’t accidentally incur costs from zombie resources and that your team knows the new landscape.

Tips for a Smooth Transition

• Start early – AWS often gives 12 months’ notice, but don’t procrastinate. Begin impact assessment within two weeks of the announcement.
• Engage AWS Support if you hit technical blockers. Some enterprise customers qualify for migration assistance credits.
• Treat discontinuation as an opportunity to modernize. A retiring service might be replaced by a newer, more efficient alternative that reduces operational costs.
• Maintain a service inventory using AWS Config and Resource Explorer. This makes future assessments faster.
• Communicate openly with your team and end users about upcoming changes. Surprise shutdowns erode trust.
• Consider open‑source alternatives where AWS lacks a direct replacement – this can also reduce vendor lock‑in.

By following these steps, you’ll turn an unexpected AWS sunset into a controlled, planned evolution of your infrastructure. The key is vigilance, prompt action, and thorough testing—qualities that will serve you beyond any single service retirement.