How do I troubleshoot 503 errors that I receive when I use a Classic Load Balancer?

Resolution

When you receive an HTTP 503 error, either the load balancer or a registered instance is causing the issue.

To view whether the hosts of your backend instances for your load balancer are healthy or unhealthy, check the HealthyHostCount and UnHealthyHostCount CloudWatch metrics. If the metrics show that your instances are running but with unhealthy hosts, then perform the following checks.

Check that your backend instances can respond to health checks

Make sure that your Classic Load Balancer can receive a 200 response code from the backend instance. For Layer 4 health checks, the load balancer marks the backend instance as healthy if the instance completes a TCP handshake. For more information, see Troubleshoot a Classic Load Balancer: Health checks.

Check that your load balancer and backend instances can handle the load

Check the SurgeQueueLength and SpilloverCount metrics to see if your instance is experiencing capacity issues.

Check the CPUUtilization CloudWatch metrics for your backend instances. If the CPU utilization is spiking to 100% or is consistently high, then add more backend instances or resize your current instances to larger sizes.

Related information

Monitor your Classic Load Balancer

Troubleshoot a Classic Load Balancer: HTTP errors