Configure weighted load balancing

This guide provides instructions for creating a weighted external passthrough Network Load Balancer deployment for each virtual machine (VM) instance using a regional backend service.

In this tutorial, you create an instance group with three VM instances and assign weights for each instance. You create an HTTP health check to report backend instance weights. Weighted load balancing is enabled on the backend service with locality load balancer policy as WEIGHTED_MAGLEV.

Before you begin

Read the Backend service-based external passthrough Network Load Balancer overview.
Install the Google Cloud CLI. For a complete overview of the tool, see the gcloud CLI overview. You can find commands related to load balancing in the API and gcloud CLI reference.

If you haven't run the Google Cloud CLI previously, first run gcloud init to authenticate.

Enable the compute API.

gcloud services enable compute.googleapis.com

Create VPC network, subnets, and firewall rules

Create a VPC network, subnet, and ingress allow firewall rules to allow connections to the backend VMs of your load balancer.

Create a VPC network and subnet.

a. To create the VPC network, run the gcloud compute networks create command:
```
gcloud compute networks create NETWORK_NAME --subnet-mode custom
```
b. In this example, the subnet's primary IPv4 address range is 10.10.0.0/24. To create the subnet, run the gcloud compute networks subnets create command:
```
gcloud compute networks subnets create SUBNET_NAME \
  --network=NETWORK_NAME \
  --range=10.10.0.0/24 \
  --region=us-central1
```
Replace the following:
- NETWORK_NAME: the name of the VPC network to create.
- SUBNET_NAME: the name of the subnetwork to create.
Create an ingress allow firewall rule to allow packets sent to destination TCP ports 80 and 443 to be delivered to the backend VMs. In this example, firewall rule allows connections from any source IP address. The firewall rule applies to VMs with the network tag network-lb-tag.

To create the firewall rule, run the gcloud compute firewall-rules create command:
```
gcloud compute firewall-rules create FIREWALL_RULE_NAME \
   --direction=INGRESS \
   --priority=1000 \
   --network=NETWORK_NAME \
   --action=ALLOW \
   --rules=tcp:80,tcp:443 \
   --source-ranges=0.0.0.0/0 \
   --target-tags=network-lb-tag
```
Replace FIREWALL_RULE_NAME with the name of the firewall rule to create.

Create VM instances and assign weights

Create three VM instances and assign weights:

Configure three backend VM instances to return the weights in the X-Load-Balancing-Endpoint-Weight header with HTTP responses. For this tutorial, you configure one backend instance to report a weight of zero, a second backend instance to report a weight of 100, and a third backend instance to report a weight of 900.

To create the instances, run the gcloud compute instances create command:

gcloud compute instances create instance-0 \
  --zone=us-central1-a \
  --tags=network-lb-tag \
  --image-family=debian-12 \
  --image-project=debian-cloud \
  --subnet=SUBNET_NAME \
  --metadata=load-balancing-weight=0,startup-script='#! /bin/bash
  apt-get update
  apt-get install apache2 -y
  ln -sr /etc/apache2/mods-available/headers.load /etc/apache2/mods-enabled/headers.load
  vm_hostname="$(curl -H "Metadata-Flavor:Google" \
  http://169.254.169.254/computeMetadata/v1/instance/name)"
  echo "Page served from: $vm_hostname" | \
  tee /var/www/html/index.html
  lb_weight="$(curl -H "Metadata-Flavor:Google" \
  http://169.254.169.254/computeMetadata/v1/instance/attributes/load-balancing-weight)"
  echo "Header set X-Load-Balancing-Endpoint-Weight \"$lb_weight\"" | \
  tee /etc/apache2/conf-enabled/headers.conf
  systemctl restart apache2'

gcloud compute instances create instance-100 \
  --zone=us-central1-a \
  --tags=network-lb-tag \
  --image-family=debian-12 \
  --image-project=debian-cloud \
  --subnet=SUBNET_NAME \
  --metadata=load-balancing-weight=100,startup-script='#! /bin/bash
  apt-get update
  apt-get install apache2 -y
  ln -sr /etc/apache2/mods-available/headers.load /etc/apache2/mods-enabled/headers.load
  vm_hostname="$(curl -H "Metadata-Flavor:Google" \
  http://169.254.169.254/computeMetadata/v1/instance/name)"
  echo "Page served from: $vm_hostname" | \
  tee /var/www/html/index.html
  lb_weight="$(curl -H "Metadata-Flavor:Google" \
  http://169.254.169.254/computeMetadata/v1/instance/attributes/load-balancing-weight)"
  echo "Header set X-Load-Balancing-Endpoint-Weight \"$lb_weight\"" | \
  tee /etc/apache2/conf-enabled/headers.conf
  systemctl restart apache2'

gcloud compute instances create instance-900 \
  --zone=us-central1-a \
  --tags=network-lb-tag \
  --image-family=debian-12 \
  --image-project=debian-cloud \
  --subnet=SUBNET_NAME \
  --metadata=load-balancing-weight=900,startup-script='#! /bin/bash
    apt-get update
    apt-get install apache2 -y
    ln -sr /etc/apache2/mods-available/headers.load /etc/apache2/mods-enabled/headers.load
    vm_hostname="$(curl -H "Metadata-Flavor:Google" \
    http://169.254.169.254/computeMetadata/v1/instance/name)"
    echo "Page served from: $vm_hostname" | \
    tee /var/www/html/index.html
    lb_weight="$(curl -H "Metadata-Flavor:Google" \
    http://169.254.169.254/computeMetadata/v1/instance/attributes/load-balancing-weight)"
    echo "Header set X-Load-Balancing-Endpoint-Weight \"$lb_weight\"" | \
    tee /etc/apache2/conf-enabled/headers.conf
    systemctl restart apache2'

Create an instance group

In this tutorial, you provide instructions to create an unmanaged instance group containing all three VM instances(instance-0, instance-100, and instance-900).

To create the instance group, run the gcloud compute instance-groups unmanaged create command:

gcloud compute instance-groups unmanaged create INSTANCE_GROUP \     
  --zone=us-central1-a

gcloud compute instance-groups unmanaged add-instances INSTANCE_GROUP \
  --zone=us-central1-a \
  --instances=instance-0,instance-100,instance-900

Replace INSTANCE_GROUP with the name of the instance group to create.

Create an HTTP health check

In this tutorial, you provide instructions to create an HTTP health check to read the HTTP response containing the backend VM's weight."

To create the HTTP health check, run the gcloud compute health-checks create command:
```
gcloud compute health-checks create http HTTP_HEALTH_CHECK_NAME \
  --region=us-central1
```
Replace HTTP_HEALTH_CHECK_NAME with the name of the HTTP health check to create.

Create a backend service

The following example provides instructions to create a regional external backend service configured to use weighted load balancing.

Create a backend service with the HTTP health check and set the locality load balancer policy to WEIGHTED_MAGLEV.
- To create the backend service, run the gcloud compute backend-services create command:
```
gcloud compute backend-services create BACKEND_SERVICE_NAME \
  --load-balancing-scheme=external \
  --protocol=tcp \
  --region=us-central1 \
  --health-checks=HTTP_HEALTH_CHECK_NAME \
  --health-checks-region=us-central1 \
  --locality-lb-policy=WEIGHTED_MAGLEV
```
  Replace BACKEND_SERVICE_NAME with the name of the backend service to create.

Add the instance group to the backend service.

To add the instance group, run the gcloud compute backend-services add-backend command:

gcloud compute backend-services add-backend BACKEND_SERVICE_NAME \
  --instance-group=INSTANCE_GROUP \
  --instance-group-zone=us-central1-a \
  --region=us-central1

Reserve a regional external IP address for the load balancer.
- To reserve one or more IP addresses, run the gcloud compute addresses create command:
```
gcloud compute addresses create ADDRESS_NAME \
 --region us-central1
```
  Replace ADDRESS_NAME with the name of the IP address to create.
  
  Use the compute addresses describe command to view the result. Note the reserved static external IP address (IP_ADDRESS).
```
gcloud compute addresses describe ADDRESS_NAME
```
Create a forwarding rule using the reserved regional external IP address IP_ADDRESS. Connect the forwarding rule to the backend service.
- To create the forwarding rule, run the gcloud compute forwarding-rules create command:
```
gcloud compute forwarding-rules create FORWARDING_RULE \
  --region=us-central1 \
  --ports=80 \
  --address=IP_ADDRESS \
  --backend-service=BACKEND_SERVICE_NAME
```
  Replace the following:
  
  FORWARDING_RULE: the name of the forwarding rule to create.
  
  IP_ADDRESS: the IP address to assign to the instance. Use the reserved static external IP address, not the address name.

Verify backend weights using backend service API

Verify that the backend weights are properly reported to the HTTP health check.

To get backend weights (along with health statuses) from a backend service, run the gcloud compute backend-services get-health command:
```
gcloud compute backend-services get-health BACKEND_SERVICE_NAME \
  --region=us-central1
```

The output is the following:

backend: https://www.googleapis.com/compute/projects/project-name/{project}/zones/us-central1-a/instanceGroups/{instance-group-name}
status:
  healthStatus:
  - forwardingRule: https://www.googleapis.com/compute/projects/{project}/regions/us-central1/forwardingRules/{firewall-rule-name}
    forwardingRuleIp: 34.135.46.66
    healthState: HEALTHY
    instance: https://www.googleapis.com/compute/projects/{project}/zones/us-central1-a/instances/instance-0
    ipAddress: 10.10.0.5
    port: 80
    weight: '0'
  - forwardingRule: https://www.googleapis.com/compute/projects/{project}/regions/us-central1/forwardingRules/{firewall-rule-name}
    forwardingRuleIp: 34.135.46.66
    healthState: HEALTHY
    instance: https://www.googleapis.com/compute/projects/{project}/zones/us-central1-a/instances/instance-100
    ipAddress: 10.10.0.6
    port: 80
    weight: '100'
  - forwardingRule: https://www.googleapis.com/compute/projects/{project}/regions/us-central1/forwardingRules/{firewall-rule-name}
    forwardingRuleIp: 34.135.46.66
    healthState: HEALTHY
    instance: https://www.googleapis.com/compute/projects/{project}/zones/us-central1-a/instances/instance-900
    ipAddress: 10.10.0.7
    port: 80
    weight: '900'
  kind: compute#backendServiceGroupHealth