GCP load balancers

This page describes how Abiquo supports load balancers in the Google Cloud Platform.
Abiquo supports regional and global load balancers, with separate sections in the user interface.
Abiquo supports external and internal load balancers on Google's VPC networks.
The supported load balancer backends include unmanaged instance groups and GCE network endpoint groups.

 

Abiquo supports load balancers in Google Cloud Platform (GCP). For more details about GCP load balancers, see https://cloud.google.com/load-balancing/docs/features

In GCP, load balancing entities can be regional or global.

In the myCloud view, there is a new Load balancers page in the Global section to manage global entities.

This is in addition to the page in the Locations section to manage regional entities. 

 GCP load balancers in the Global section of myCloud view
GCP load balancers in the Global section of myCloud view

 

To use a load balancer in GCP, first create a target group. Depending on the type of target group you select, different UI fields will display.

NOTE: You cannot create target groups in Abiquo of type: MANAGED_INSTANCE_GROUP and REGIONAL_MANAGED_INSTANCE_GROUP. However, you can onboard these target groups into Abiquo.

Here are some examples of typical target groups

  • Unmanaged instance group

    • Note that you should enter a Name for the receiving Port, which you will also need to enter in the Load balancer's Routing rule tab

    • And select the Global networkSubnetLocation, and Availability zone of the VMs that you will add to the target group

       Create an unmanaged instance group
      Create an unmanaged instance group

       

  • NetworkEndpointType - GCE IP port

    • RECOMMENDED

    • With this target group, you can load balance addresses and ports in Google Cloud Platform networks 

  • Note: This screenshot shows Internet IP port, which only works with Classic load balancers 

After you create the target group, you can create a load balancer, in the Location (regional) or Global section.

Abiquo supports load balancers that are internal (using a subnet IP address) and external (using a public IP address). You can select one IP address for your load balancer. If you are creating an internal load balancer with a subnet address, you will need to select a private subnet.

The Algorithm in Abiquo represents the session affinity in GCP. See https://cloud.google.com/load-balancing/docs/backend-service#session_affinity

For the Routing rules, if your target group is an unmanaged instance group, then you will need to supply the same name of the port from the target group as the "out name".

In GCP, Abiquo Routing rules represent the Forwarding rule protocol and port + target proxy (if proxy based) + URL map (if HTTP(s)) + backend service.

GCP

Abiquo

GCP

Abiquo

Forward rule protocol, port

Routing rule protocol, Port In

Backend service protocol, named port

Routing rule protocol OUT, port name out (for proxy based load balancers)

Backend service backends

Forward action, target groups target of the default conditional action:

Balancing mode:

  • RATE: HTTP, HTTPS, HTTP2

  • CONNECTION: SSL, TCP, UDP

See https://cloud.google.com/load-balancing/docs/backend-service#balancing-mode

Configure the routing rules of your load balancer in GCP.

You will need to create a health check for each backend service.

After you create a load balancer, you can add nodes to the target groups.

The nodes may be VMs.

Or the nodes may be IP addresses and ports.

You can check the health of the VM nodes on the Target Group details panel. After you select the Load balancer from the drop-down in the top-right hand corner, the platform will display the Status of each node.

To add or change a target group, edit the load balancer, go to Routing rules, select the routing rule, and click Edit conditional actions.

To edit the conditional action, click the pencil edit button, make your changes, and click Accept. When you finish editing, save the routing rule with the load balancer.

The Weight represents the maximum connections per instance/endpoint for CONNECTION protocols, and the maximum rate per instance/endpoint for RATE protocols. See https://cloud.google.com/load-balancing/docs/backend-service#target_capacity .

 

Copyright © 2006-2024, Abiquo Holdings SL. All rights reserved