413 Payload Too Large Error in ALB Ingress Controller

[shivraj001]

Bug Description

When sending an HTTP request via ALB Ingress, we encounter a 413 Payload Too Large error. The issue persists even after setting the idle connection timeout to 300 seconds.

Steps to Reproduce

1. Deploy an application behind an AWS ALB Ingress Controller.
2. Configure the ingress resource with an annotation to increase the idle timeout:

   metadata:
     annotations:
       alb.ingress.kubernetes.io/load-balancer-attributes: idle_timeout.timeout_seconds=300

3. Attempt to send a large HTTP request (greater than 50MB) to the service via ALB.
4. Observe the 413 Payload Too Large error in response.

Expected Behavior

The ALB should accept large payloads as expected and pass the request to the backend service without a 413 error.

Actual Behavior

• The request fails with a 413 Payload Too Large error.
• Increasing the idle_timeout did not resolve the issue.
• Logs from ALB Ingress Controller do not show relevant errors but the client request is rejected.

Regression

Was the functionality working in a previous version? No

Current Workarounds

• No definitive workaround found.
• Attempted increasing idle timeout but issue persists.

Environment

• AWS Load Balancer Controller Version: v2.10.0
• Kubernetes Version: 1.31
• Using EKS: Yes
• Using Service or Ingress: Ingress
• ALB Ingress Annotations:

  alb.ingress.kubernetes.io/load-balancer-attributes: idle_timeout.timeout_seconds=300
  ```  .  
  

[zac-nixon]

Have you checked to make sure the backend webserver will allow 50mb payloads?

[zac-nixon]

Please also check your header sizes: https://docs.aws.amazon.com/elasticloadbalancing/latest/application/load-balancer-troubleshooting.html#http-413-issues

ALB should not enforce any payload size limitations (as long as you aren't using Lambda targets)