DevelopersGlossarySign Up

Create a Load Balancer Instance

A Load Balancer distributes incoming requests or network traffic among a set of backend servers based on predefined algorithms and rules. Load balancers also perform health checks on backend servers to ensure that they are capable of handling requests.

The primary purpose of a load balancer is to optimize resource utilization, maximize throughput, minimize response time, and ensure high availability of applications or services.

Procedures

Go to Elastic Compute > Load Balancer > Create Load Balancer.

1

Choose a pricing model

Currently only pay-as-you-go is supported.

2

Select a location

Select the location where you want to have your load balancer.

Please note that your load balancer and your backend servers should locate in the same region.

3

Select the specification

Currently the Standard model is available with unlimited max connections, CPS and QPS.

  • Max connections: Max Connections refer to the maximum number of connections that a load balancing instance can handle.

  • CPS: Connections Per Second (CPS) refers to the number of new connections that a load balancing instance can handle in one second.

  • QPS: Queries Per Second (QPS) refers to the rate at which the load balancing instance can handle incoming queries or requests in one second, specifically in Layer 7 (application layer) listening.

4

Select global VPC for the load balancer

The load balancer must be deployed within a specific global VPC and can only be associated with backend servers within this VPC. Therefore, the backend server’s network interface (vNIC) must be in the same global VPC as the load balancer.

If no desired global VPC is available, click New Global VPC to create one.

5

(Optional) Associate security group

You can associate your load balancer to a security group whenever you want to control or restrict inbound and outbound traffic to the load balancer.

6

Select IP stack type and default elastic IPv4

The elastic public IPv4 configured here serves as the VIP (Virtual IP), which is the external access address of the load balancer.

IP stack type specifies the IP protocol version supported by the load balancer. Currently, only IPv4 is supported, meaning the load balancer is assigned an IPv4 VIP only and accepts traffic from IPv4 clients only.

  • BGP IP: Utilizes Border Gateway Protocol for flexible and reliable routing, suitable for highly available network scenarios. BGP IP provides optimized multi-carrier routing.

  • Local IP: Sourced from a local ISP and matches the country of the data center, commonly used for regulatory compliance or local presence. Local IP is suitable for regional or cost-sensitive scenarios.

You can also assign elastic IPv4 later after creation. See Bind Elastic IPv4 for details.

7

Select public network billing method

Public internet traffic forwarded by the load balancer incurs charges. Please select an appropriate public network billing method based on your business requirements. Currently, elastic public IPs support Data Transfer, Flat Rate, and Bandwidth Cluster billing models. If you would like to use Bandwidth Cluster billing, please submit a support ticket to contact us.

  • Data Transfer Configure a data transfer package. Any traffic exceeding the package limit will incur additional charges. You can set the upper limit of public network bandwidth to control traffic usage in case of extra high charge. If you don’t set the bandwidth cap, it will be the default value.

  • Flat Rate Configure a fixed bandwidth cap. It is suitable for businesses with stable and predictable public network usage.

  • Bandwidth Cluster Share a bandwidth commitment across multiple compute resources (including load balancers) in a region or city. Billing is based on the 95th percentile of peak bandwidth usage, offering flexibility for burst traffic and cost optimization if you have unpredictable or periodic high demand.

8

Label your load balancer

Give your load balancer an identical label name.

The name must be 2 to 63 characters long and can contain letters, numbers, hyphens (-), underscores (_), slashes (/), and periods (.). It must start and end with a letter or number.

9

Custom health probe private source IP

Select 2 private IPv4 addresses from a same-region subnet for health checks. If not set, they will be randomly assigned from any same-region subnet in the global VPC.

If no desired subnets in the same region, click New Subnet to create one.

10

More settings

Manage resources

You can add your load balancer into a specified resource group. In a resource group, specific members can be granted defined permissions for resources.

See Create a Resource Group for more details.

Set tags / Edit tags (On the load balancer list page)

Tags are metadata labels that you can assign to resources.

See Add Tags to Resources for more details.

Result

After creation, the load balancer enters the Active state. On load balancer list page, you can view the load balancer’s state, name and ID, region, IP type, network configuration, and whether listeners have been configured.

What to Do Next

After creation, click the load balancer name to go to Details page. Go to Information to view specification, listener list, network configuration, order and billing, and etc.

You must configure listeners and backend servers for a load balancer to become fully operational.

See add listeners to the load balancer and add backend servers to a listener for more details.

You can also perform the follow operations if you need:

PreviousZenlayer Load Balancer OverviewNextCheck Load Balancer Performance

Last updated