I’m getting 504 status code returns when requesting my load balancer, it then transfers the request to a fargate container.
This happens when launching simultaneous calls, some return 200, some 504.
I tried raising the Idle timeout to 60 seconds in the UI as well as in the aws cli but my requests keep timing out after 10 seconds.
Is this normal behavior ? Is their another way to raise this threshold ?
Thanks
2
Answers
Seems more a connection issue between the ALB and some of the instances, that causes an intermittent Gateway Timeout.
Check if the subnets where the instances are deployed are equally configured and, if you use a Network Load Balancer, enable Cross-Zone Load Balancing.
504 Gateway Timeout means that the ALB couldn’t reach the backend, this is also quoted here: "The load balancer failed to establish a connection to the target before the connection timeout expired (10 seconds)."
Raising the timeout (which is an idle connection timeout) has no effect. Please try to troubleshoot the connection between the ALB and the Fargate Container.
The first and best check is to see if the TargetGroup is healthy.