Delete the load balancer and backend services. Create a new passthrough Network Load Balancer. Configure a failover group of VMs for the backend.

Access a VM in the VPC through SSH and try to access a backend VM directly. If the request is successful from the VM, increase the quantity of backends.

Enable and review the health check logs. Review the error responses in Cloud Logging.

Validate the health of the backend service. Enable logging for the backend service and identify the error response in Cloud Logging. Determine the cause of the error by reviewing the statusDetails log field.

Validate the health of the backend service. Enable logging on the load balancer and identify the error response in Cloud Logging. Determine the cause of the error by reviewing the statusDetails log field.