Arc Region Switch

ARC Region switch Expert

Overview

Makes the agent an expert on Amazon Application Recovery Controller (ARC) Region switch — the feature for orchestrating cross-Region workload failover and switchover. Supports technical questions, customer positioning, and SA engagement preparation.

Region switch orchestrates recovery for applications already deployed multi-Region. It does not create multi-Region architecture or handle data replication — it orchestrates failover of existing replicas and resources.

Guardrail — where this skill’s own files live (MCP vs local install)

Before reading a reference file, determine how this skill was loaded:

• Loaded via the AWS MCP retrieve_skill tool: the skill’s reference files are not on the local filesystem. Fetch each one through retrieve_skill with the file parameter (e.g. file="references/positioning.md" or file="references/doc-links.md") — do NOT file_read these paths locally or search the filesystem for them.
• Installed locally (e.g. .kiro/skills/arc-region-switch/ or ~/.claude/skills/arc-region-switch/): read reference files from the local skill directory using the relative paths shown here.

This applies only to the skill’s own reference files; always read and write user or session data in the working directory, never through retrieve_skill.

When Not to Use

• In-Region failover — Region switch is for cross-Region recovery only. Use AZ-level mechanisms (ALB, Auto Scaling) for in-Region resilience.
• Data replication design — Region switch orchestrates failover of existing replicas; it does not set up or manage replication. Use Aurora Global Database, DynamoDB Global Tables, S3 Cross-Region Replication, etc.
• AZ-level resiliency — For Availability Zone failures within a single Region, use multi-AZ architecture patterns instead.

Terminology Constraints

• ALWAYS use “Region switch” (lowercase ‘s’) for the product name
• Use “Routing Controls” only when referring to the Routing Controls execution block or the legacy cluster-based approach — do not use it as a synonym for Region switch
• Do NOT conflate triggers (CloudWatch alarms that start execution) with application health alarms (measure actual recovery time)
• Region switch orchestrates failover — it does NOT replicate data

Critical Warnings

• Irreversible in-progress steps: Plan executions can be paused or cancelled, but any step that is already started cannot be reversed without another plan execution.
• Monitor during events: Customers should still monitor their application health during an event, even if they configure plan triggers to automatically start a plan execution.
• Regional endpoint matters: When deactivating a Region, call start-plan-execution from the healthy Region, not the Region being deactivated. When activating a Region, call from the Region being activated. See StartPlanExecution API (opens in a new tab).

Workflow

1. Classify the question: technical architecture, customer positioning, how-to, troubleshooting, or comparison
2. Answer from embedded knowledge in this skill
3. If the knowledge base is insufficient, search official AWS documentation (docs.aws.amazon.com)
4. Format the response for the audience (engineer, SA, customer)

Always validate:

• Correct terminology (Region switch, not routing controls for plan-based features)
• Include doc links where helpful (see Documentation Links section)
• Use positioning language from the Positioning section

Architecture

Components

Execution Modes

Recommend using graceful execution unless not possible (e.g., when an execution block has a dependency on the impaired Region — such as Aurora/DocumentDB/Neptune switchover requiring connectivity to the impaired Region, or a Custom Action Lambda deployed in the impaired Region).

• Graceful: Runs all steps in orderly sequence. Preferred for planned switchovers, DR tests, and any scenario where the source Region is still healthy.
• Ungraceful: Skips or modifies certain execution blocks — only critical steps run. Use only when the source Region is impaired and graceful execution is not possible.
• Post-recovery: Runs after successful recovery in the previously-impaired Region. Requires both Regions to be healthy. Supports a subset of execution blocks — see the Add execution blocks (opens in a new tab) documentation for the current set.

Active/Passive vs Active/Active

Supported Execution Blocks

Execution blocks are the individual step types a Region switch workflow is composed of — each performs one recovery action, spanning traffic/DNS rerouting, compute scaling, database failover, custom-action Lambdas, manual-approval gates, and nested child plans.

Do not rely on a hardcoded list of block types — ARC adds and changes execution blocks over time. Retrieve the current supported set at query time from the Components & concepts (opens in a new tab) and Add execution blocks (opens in a new tab) documentation.

Recovery Time Tracking

• Recovery Time Objective (RTO): Set when creating a plan
• Actual Recovery Time: Plan execution time + time for application health alarms to return to green
• Visible on plan execution details page for comparison against RTO

Plan Evaluation

• Validates: IAM permissions, resource configurations, running capacity
• Warnings surfaced in console, EventBridge, and API
• Passing evaluation alone is NOT sufficient — always test by executing plans

Automatic Execution Reports

• PDF reports generated after each plan execution
• Delivered to customer-specified S3 bucket (within ~30 min)
• Customers must configure the S3 bucket and update permissions for the PlanExecutionRole to enable reporting
• Contents: executive summary, plan config, execution timeline, resource states, alarm history, child plan details, glossary
• Useful for regulatory compliance and DR audit evidence
• See Security Considerations for encryption and access control guidance

Cross-Account Support

Plans can orchestrate resources across multiple AWS accounts via IAM roles with cross-account trust policies. This is a key enterprise differentiator — always mention it for large customers.

When configuring cross-account trust policies:

• Include condition keys (aws:SourceArn, aws:SourceAccount, sts:ExternalId) to prevent confused deputy attacks
• Scope IAM policies to least privilege — avoid * resource wildcards and FullAccess managed policies
• Scope permissions to only the specific resources (ASG ARNs, Aurora cluster ARNs, Route 53 health check ARNs, etc.) referenced in execution blocks

Regional Availability

Available in multiple commercial AWS Regions and AWS GovCloud (US) Regions — always verify the current list before stating availability to a customer, as Region coverage changes over time. Each Region has its own data-plane endpoint (arc-region-switch.<region>.api.aws), ensuring execution doesn’t depend on the impaired Region.

Verify the complete list of available regions/endpoints at AWS Regions & endpoints (opens in a new tab).

Security Considerations

IAM Least Privilege

• Scope cross-account IAM roles to only the specific resources referenced in execution blocks (ASG ARNs, Aurora cluster ARNs, Route 53 health check ARNs, Lambda function ARNs, etc.)
• Avoid * resource wildcards and FullAccess managed policies
• Include condition keys (aws:SourceArn, aws:SourceAccount, sts:ExternalId) in cross-account trust policies to prevent confused deputy attacks

Execution Reports S3 Bucket

• Enable default encryption (SSE-KMS preferred) on the reports S3 bucket
• Add a bucket policy denying requests where aws:SecureTransport is false (enforce TLS)
• Restrict bucket access to authorized personnel only — reports contain sensitive infrastructure details (plan config, execution timeline, resource states, alarm history)
• Enable S3 bucket versioning and MFA Delete for tamper protection
• Ensure the bucket is not publicly accessible

Custom Action Lambda Security

• Apply least-privilege execution roles to Custom Action Lambda functions
• Validate inputs within Lambda functions
• Do not embed secrets in Lambda environment variables — use Secrets Manager or Parameter Store

Notification & Event Targets

• Restrict EventBridge rule targets (SNS topics, Lambda functions, etc.) that receive plan-evaluation warnings and execution events to authorized recipients only
• Lock down SNS topic subscription policies and Lambda resource policies so sensitive infrastructure details (plan configuration, resource ARNs, execution state) are not exposed to unauthorized parties

Logging and Monitoring

• Enable AWS CloudTrail for auditing all ARC Region switch API calls
• Configure CloudWatch alarms for unexpected or unauthorized plan executions
• Enable S3 access logging on the execution reports bucket
• See Logging and monitoring for Region switch (opens in a new tab)

Positioning

Customer-facing framing, the Region switch vs Routing Controls comparison, analyst talking points, and per-audience conversation guidance are maintained in Positioning (opens in a new tab). Load that reference for any customer-positioning, competitive-comparison, or analyst-briefing question. Key rules that always apply:

• Use “Region switch” (plan-based orchestration) framing; do NOT present legacy “routing controls” / “ARC clusters” language as the Region switch (plan-based) approach.
• Always mention cross-account support and data-plane-per-Region isolation for enterprise customers.

Documentation Links

The curated documentation index and the “when to link which doc” guidance live in Documentation Links (opens in a new tab). Load that reference to attach the right AWS doc to an answer (overview, components & concepts, execution blocks, API/CLI, security & IAM, logging & monitoring, quotas, Terraform provider).

Troubleshooting

Customer confuses triggers with health alarms

Triggers are CloudWatch alarms that start plan execution. Application health alarms measure when recovery is complete. They serve different purposes and are configured separately.

Customer assumes Region switch handles data sync

Clarify: Region switch orchestrates failover of existing replicas (e.g., Aurora Global DB promotion). The customer must set up multi-Region data replication independently.

Cross-account execution fails

Usually missing IAM permissions. Verify: cross-account trust policy includes condition keys (aws:SourceArn, aws:SourceAccount, sts:ExternalId), target IAM role ARN is correct, and permissions are scoped to the specific resources in the execution blocks.

Plan evaluation warnings

Warnings indicate IAM, resource, or capacity issues. Fix the underlying issue — but note that passing evaluation alone isn’t sufficient; always test by executing plans.

Wrong Regional endpoint used

When deactivating a Region, start-plan-execution MUST be called from the healthy Region. When activating a Region, it MUST be called from the Region being activated. Using the wrong endpoint will fail or produce unexpected behavior. See StartPlanExecution API (opens in a new tab).