Skip to content

AWS Global Accelerator - Runbook & Reference

中文 · GitHub source Facts verified against official AWS documentation: 2026-08-19

Overview

AWS Global Accelerator improves the availability and performance of internet applications for a global audience. It provides static anycast IP addresses and routes traffic over the AWS global network to the optimal regional endpoint based on health, client location, and your policies.

Key concepts

  • Accelerator: the global resource that directs traffic; provides two static IPv4 addresses (or four for dual-stack) that stay assigned for the accelerator's lifetime.
  • Listener: receives traffic on specific ports/protocols (TCP/UDP) and routes to endpoint groups.
  • Endpoint group: a regional group of endpoints; traffic shifts based on health and weights.
  • Endpoints: NLB, ALB, EC2 instances, or Elastic IPs in one or more Regions.
  • Custom routing accelerator: maps users to specific destinations (VPC subnet private IPs) instead of load-balanced endpoints.
  • Health checks: Global Accelerator reacts instantly to endpoint health changes; respects Application Recovery Controller zonal shifts.

Common operations (AWS CLI)

# Create an accelerator
aws globalaccelerator create-accelerator --name prod --ip-address-type IPV4

# Add a listener (443/TCP)
aws globalaccelerator create-listener --accelerator-arn <accelerator-arn> \
  --port-ranges From=443,To=443 --protocol TCP

# Create an endpoint group with an ALB endpoint
aws globalaccelerator create-endpoint-group --listener-arn <listener-arn> \
  --endpoint-group-region us-east-1 \
  --endpoint-configurations EndpointId=arn:aws:elasticloadbalancing:us-east-1:123456789012:loadbalancer/app/web/1234567890abcdef,Weight=100

# List and inspect
aws globalaccelerator list-accelerators
aws globalaccelerator describe-accelerator --accelerator-arn <accelerator-arn>
aws globalaccelerator list-listeners --accelerator-arn <accelerator-arn>

Best practices

  • Use Global Accelerator for global, latency-sensitive, or availability-critical applications instead of DNS-based failover only.
  • Put ALB/NLB behind it and keep DNS TTLs short; update endpoints in the accelerator when infrastructure changes.
  • Configure endpoint groups in multiple Regions with weights for active/passive or active/active routing.
  • Use health checks and test failover behavior; respect VPC Block Public Access settings.
  • Protect accelerator deletion with IAM/tag policies; the static IPs are lost if the accelerator is deleted.
  • Use custom routing accelerators for gaming/real-time apps that need affinity to a specific destination.

Troubleshooting

Symptom Checks and fixes
Clients can't connect Check listener ports/protocol, endpoint group health, and that endpoints accept traffic.
Traffic to unhealthy Region Review endpoint health status and weights; confirm health checks pass.
Static IPs lost Static IPs are released only on accelerator deletion; avoid deleting production accelerators.
Endpoint unreachable Verify security groups/NACLs allow traffic from Global Accelerator (use the AWS published ranges).
Performance not improved Confirm DNS points to the accelerator static IPs and clients reach the nearest edge.

Limits

Accelerators per account, listeners, endpoint groups, and endpoints have quotas. See the Service Quotas console for current values.

Official references