Skip to content Skip to footer

What AWS’s New AI Configuration, Migration and Operations Features Mean for Production Teams

What Happened

On November 19, 2026, AWS announced updates spanning AI applications, migrations, databases, operations and Marketplace workflows. These are feature releases, not new service versions; AWS did not specify version numbers for them.

  • AI applications: AWS AppConfig added separate profiles for model invocation settings and prompts, with gradual rollout and alarm-triggered rollback. The AWS Lambda Durable Execution SDK for Rust became generally available, supporting checkpointed workflows that can wait for external events for up to one year. [22] [14]
  • Migrations: AWS Transform combined assessment and planning in one job, added application-level discovery and tracking, recommended EBS and FSx for NetApp ONTAP storage configurations, and introduced reusable actions for preparation and cutover. [2] [3] [11] [21]
  • Data and operations: Aurora DSQL added cross-continent multi-Region peering; MSK Connect added in-place connector and task restarts. Higher Parameter Store throughput quotas now apply to five operations when the account’s higher-throughput setting is enabled. EC2 added performance status checks and optional, more granular Detailed Monitoring. AWS Health also announced tag-enriched events, but did not provide a launch date for that capability. [8] [7] [10] [19] [16]
  • Partners and purchasing: AWS Revenue Measurement added deal-level attribution, closed-source Bedrock model coverage, and SaaS account and IAM identity tagging. Partner Central added AI-assisted listings; Marketplace added buyer contact sharing and purchasing approvals. [1] [12] [13] [17] [6] [9] [18]
  • Other application features: Amazon IVS released a Linux broadcast SDK with a C API and Go bindings. MediaTailor added Dynamic Avail Sizing for live HLS. New Amazon Connect Customer instances now enable Cases by default, while supported WorkSpaces Personal users can enable administrator-permitted Nested Virtualization. [4] [5] [15] [20]

Why It Matters to Businesses

The most immediate engineering benefit is operational control: teams can change and roll back AI prompts without redeploying, restart failed streaming connectors without recreating them, and plan migration costs and waves against the same inventory. These reduce manual steps, but do not remove the need to validate application behavior, data dependencies or cutover readiness. [22] [7] [2] [3]

For buyers and sellers, the Marketplace changes affect procurement and onboarding as much as technology. Approval policies can route purchases through existing systems; sellers can receive buyer contact details when buyers choose to share them. Revenue attribution gives partners more ways to connect product usage to deals, subject to the tagging and account-dedication rules of each method. [18] [9] [1] [13]

Kimbodo Engineering Perspective

Prioritize changes that improve a production control loop. AppConfig’s staged rollout and rollback are useful only when application-quality, error and token-usage metrics are trustworthy. Durable execution simplifies progress tracking, but each step still needs a clear retry and idempotency policy. For migrations, application-owner confirmation and tested cutover actions matter more than an automatically generated wave plan. [22] [14] [3] [21]

How We Would Implement It

  • Inventory current prompts and model settings, move eligible values into separate AppConfig profiles, and release changes to a small cohort first. Attach CloudWatch or existing observability alarms to rollback criteria. [22]
  • For Rust workflows, prototype one bounded process with the durable SDK. Test resumes, callbacks and duplicate side effects locally before deploying on provided.al2023. [14]
  • For a migration, upload one inventory to AWS Transform, have owners verify application boundaries, then compare storage recommendations against measured performance. Rehearse source actions and define an explicit go/no-go decision for each wave. [2] [3] [11] [21]
  • Update runbooks to use MSK Connect restarts and the applicable Parameter Store quotas. Evaluate EC2’s optional metrics only where their diagnostic value justifies the CloudWatch cost. [7] [10] [19]

Risks, Costs and Security

Prompt and model configuration changes can alter output, latency and spend without a code deployment; restrict who can publish them and retain approval and rollback records. Durable workflows need safeguards against repeated external actions after retries. Cross-continent database placement requires review of data-residency requirements, latency and failure behavior before adoption. [22] [14] [8]

Keep migration automation scoped to approved servers and waves, especially actions that stop services or freeze writes. Treat Marketplace contact sharing as customer data, and validate buyer approval policies before routing purchases. Detailed Monitoring incurs standard CloudWatch charges; the higher Parameter Store quotas are conditional, not a universal default increase. [21] [9] [18] [19] [10]

Where Kimbodo Comes In

Kimbodo builds and operates this in production for businesses — see our AI Application Development practice, or Estimate My AI Application.

Sources

  1. [1] AWS Revenue Measurement now supports deal-level revenue attribution for AWS Marketplace offers and opportunities
  2. [2] AWS Transform combines migration assessment and planning in one experience
  3. [3] AWS Transform extends application-centric migration across the lifecycle
  4. [4] Amazon IVS Real-Time Streaming now provides a Linux broadcast SDK
  5. [5] AWS Elemental MediaTailor adds Dynamic Avail Sizing for live HLS
  6. [6] AWS Partner Central now supports AI-assisted listing experience for product, services, and solutions
  7. [7] Amazon MSK Connect now supports restarting existing connectors
  8. [8] Amazon Aurora DSQL now supports multi-Region peering across continents
  9. [9] AWS Marketplace now supports buyer contact sharing at subscription
  10. [10] AWS Systems Manager Parameter Store raises API throughput limits
  11. [11] AWS Transform supports optimized storage for server migration
  12. [12] AWS Revenue Measurement now supports closed-source models in Amazon Bedrock
  13. [13] AWS Revenue Measurement now supports Account Tagging for SaaS Products
  14. [14] AWS Lambda Durable Execution SDK for Rust is now generally available
  15. [15] Amazon Connect Customer enables Cases by default for new instances
  16. [16] AWS Health now supports resource and account tags
  17. [17] AWS Revenue Measurement now supports Identity Tagging
  18. [18] AWS Marketplace now supports buyer approval workflows
  19. [19] Amazon EC2 launches performance status checks and new utilization metrics for EC2 Detailed Monitoring
  20. [20] Amazon WorkSpaces now lets developers enable Nested Virtualization
  21. [21] AWS Transform automates source and migration target actions to accelerate cutover
  22. [22] AWS AppConfig launches support for managing native AI model and AI prompt configuration

Leave a comment

0.0/5

Engineering Founded & Customer Focused

KIMBODO LLC. © 2026. All Rights Reserved.