Amazon Application Recovery Controller FAQs

Amazon Application Recovery Controller (ARC) is a suite of recovery services for your single-Region and multi-Region applications. ARC's single-Region capabilities (zonal shift and zonal autoshift) move traffic away from an impaired AZ. ARC's multi-Region capabilities (Region switch and routing controls) orchestrate full-stack failover and DNS-based traffic shifting across Regions.

ARC supports single-Region recovery for multi-AZ applications using zonal shift and zonal autoshift, and multi-Region recovery using Region switch and routing controls.

You can get started through the ARC console. For single-Region recovery, enable zonal shift or zonal autoshift on your supported resources. For multi-Region recovery, create a Region switch plan or set up routing controls.

Zonal shift and zonal autoshift are available free of cost. Region switch is priced at $70 per plan per month. Routing control clusters are priced at $2.50 per hour per cluster. See the pricing page for full details.

Region switch is a failover orchestration service that orchestrates the failover of your entire application stack from one Region to another, for active/active and active/passive workloads.

Yes. You can include application resources from multiple AWS accounts in a single recovery plan and orchestrate recovery across accounts from a single account.

Yes. Trigger plan execution automatically based on Amazon CloudWatch alarm state changes, or execute plans manually through the console or API.

Both active/passive configurations (failover and failback) and active/active configurations (shift-away and return). For active/passive, Region switch can scale up and configure your application stack, including compute and database instances, in the standby Region before shifting traffic.

Zonal shift allows you to manually shift traffic for a supported resource away from an impaired Availability Zone to healthy AZs in the same Region with a single action.

You set an initial expiration from 1 minute up to 3 days (72 hours). You can extend the expiration or cancel a zonal shift at any time.

Application Load Balancers, Network Load Balancers, EC2 Auto Scaling Groups, and Amazon EKS clusters.

Yes. Ensure your application has sufficient capacity in the remaining Availability Zones to handle the redirected traffic. Do not rely on auto scaling alone when initiating a zonal shift.

Zonal autoshift authorizes AWS to automatically shift your application's traffic away from an impaired Availability Zone on your behalf when internal telemetry detects an AZ impairment.

Mandatory weekly tests where ARC shifts traffic for a resource away from one AZ for approximately 30 minutes. Practice runs validate that your application can operate without one AZ and record an outcome (SUCCEEDED or FAILED).

Yes. Configure blocked windows to define time periods when practice runs should not occur, such as during peak traffic hours or maintenance windows.

Zonal shift is manual—you start and stop it. Zonal autoshift is automatic—AWS detects impairments and shifts traffic on your behalf. Zonal autoshift also includes mandatory practice runs.

Simple on-off switches hosted on highly available clusters spanning five AWS Regions. Integrated with Amazon Route 53 health checks, they let you reroute traffic between Regions without complex DNS manipulation.

A set of redundant regional endpoints across five AWS Regions against which you execute API calls to update or get the state of routing controls. Each account can have a maximum of 2 clusters, and each cluster can host multiple routing controls.

Guardrails you define to prevent unintended recovery actions. For example, an assertion rule can prevent all routing controls from being turned off simultaneously, or a gating rule can require approval before a failover proceeds.