First of all, we need to create a backend service that specifies health check. Then, we must add each instance group with a threshold: CPU utilization up to 80% and the maximum capacity set to 100% for both HTTP and Tomcat:

//create backend service for http (default) and named port tomcat (8080/tcp)
$ gcloud compute backend-services create my-http-backend-service --health-checks my-http-health-check --protocol HTTP --global
$ gcloud compute backend-services create my-tomcat-backend-service --health-checks my-tomcat-health-check --protocol HTTP --port-name tomcat --global


//add http instance groups (both us-west1 and us-east1) to http backend service
$ gcloud compute backend-services add-backend my-http-backend-service --instance-group http-ig-us-west --instance-group-zone us-west1-a --balancing-mode UTILIZATION --max-utilization 0.8 --capacity-scaler 1 --global
$ gcloud compute backend-services add-backend my-http-backend-service --instance-group http-ig-us-east --instance-group-zone us-east1-c --balancing-mode UTILIZATION --max-utilization 0.8 --capacity-scaler 1 --global


//also add tomcat instance group to tomcat backend service
$ gcloud compute backend-services add-backend my-tomcat-backend-service --instance-group tomcat-ig-us-west --instance-group-zone us-west1-a --balancing-mode UTILIZATION --max-utilization 0.8 --capacity-scaler 1 --global