Creating a Backend Server Group

Scenario

To route requests, you need to associate at least one backend server group with each listener.

A backend server group can be associated with listeners of different load balancers.

Table 1 describes the scenarios for creating a backend server group.

**Table 1** Scenarios
Scenario	Reference
Creating a backend server group and associating it with a load balancer	Procedure
Creating a backend server group when adding a listener	Adding listeners with different protocols by referring to Listener Overview
Changing the backend server group associated with a listener	Changing a Backend Server Group

Constraints

The backend protocol of the new backend server group must match the frontend protocol of the listener as described in Table 2.

**Table 2** The frontend and backend protocol
Load Balancer Specification	Frontend Protocol	Backend Protocol
Network load balancing	TCP	TCP
Network load balancing	UDP	UDP QUIC
Network load balancing	TLS	TLS TCP
Application load balancing	HTTP	HTTP
Application load balancing	HTTPS	HTTP HTTPS gRPC
Application load balancing	QUIC	HTTP HTTPS

Procedure

Go to the backend server group list page.
Click Create Backend Server Group in the upper right corner.

Configure the routing policy based on Table 3.

**Table 3** Parameters required for configuring a routing policy
Parameter	Description
Backend Server Group Name	Specifies the name of the backend server group.
Type	Specifies the type of load balancer that can use the backend server group. Select Dedicated.
Load Balancer	Specifies whether to associate a load balancer. You can associate an existing dedicated load balancer when you create a backend server group or associate one later. Associate later Associate existing
Forwarding Mode	Specifies the forwarding mode to distribute traffic. There are two options: Load balancing and Active/Standby. Load balancing: You can add one or more backend servers to the backend server group. Active/Standby: You must add two backend servers to the backend server group, one acting as the active server and the other as the standby server. If the active server is faulty, traffic is forwarded to the standby server, improving service reliability. Active/standby backend server groups can only be associated with TCP, UDP, and TLS listeners.
Backend Server Group Type	Specifies the type of the backend server group. Hybrid: You can add cloud servers, supplementary network interfaces as backend servers, and add IP addresses as backend servers when IP as a Backend is enabled. When you create a hybrid backend server group, you must specify a VPC and associate the backend server group with a load balancer in this VPC. IP as a backend server: You can add IP addresses as backend servers only when you enable IP as a Backend. NOTE: This option is available in certain regions. You can see which regions support this option on the console.
VPC	Specifies the VPC where the backend server group works. You can associate the backend server group with a load balancer in this VPC. This parameter is mandatory if you select Hybrid for Backend Server Group Type. You can select an existing VPC or create a new one. For more information about VPC, see the Virtual Private Cloud User Guide.
Backend Protocol	Specifies the protocol that backend servers in the backend server group use to receive requests from the listeners. The protocol varies depending on the forwarding mode: Load balancing: HTTP, HTTPS, gRPC, TCP, UDP, TLS, or QUIC Active/Standby: TCP, UDP, TLS, or QUIC
IP Address Version	Specifies the IP address version of backend servers that can be added to the backend server group. By default, IPv4 backend servers are supported. There are two options when the backend protocol is TCP or UDP: IPv4: Only IPv4 addresses can be added as backend servers. Dual stack: Both IPv4 and IPv6 addresses can be added as backend servers. NOTE: This option is available in certain regions. You can see which regions support this option on the console.
Forward to Same Port	If this option is enabled, you do not need to specify a backend port when you add a backend server. The listener routes the requests to the backend server over the same port as the frontend port. This option cannot be disabled after being enabled. NOTE: This option is available in certain regions. You can see which regions support this option on the console. NOTE: This option is available only for TCP and UDP backend server groups associated with a dedicated load balancer.
Load Balancing Algorithm	Specifies the algorithm used by the load balancer to distribute traffic. The following options are available: Weighted round robin: Requests are routed to different servers based on their weights. Backend servers with higher weights receive proportionately more requests, whereas equal-weighted servers receive the same number of requests. Weighted least connections: In addition to the number of connections, each server is assigned a weight based on its capacity. Requests are routed to the server with the lowest connections-to-weight ratio. Source IP hash: Allows requests from different clients to be routed based on source IP addresses and ensures that requests from the same client are forwarded to the same server. Connection ID: This algorithm is available when you have selected QUIC for Backend Protocol. This algorithm allows requests with the same connection ID to be routed to the same backend server. For more information about load balancing algorithms, see Load Balancing Algorithms.
Forwarding Even Unhealthy	Specifies whether to forward traffic across all the backend servers even if all of them have been identified as unhealthy. This option is only available if Forwarding Mode is set to Load balancing. This option is disabled by default, preventing requests from being sent to the backend server group, in which all backend servers are identified as unhealthy. If this option is enabled, ELB forwards traffic across all the backend servers even if all of them have been identified as unhealthy. The function improves service availability by preventing disruptions from faulty health checks resulting from misconfigurations.
Sticky Session	Specifies whether to enable sticky sessions if you have selected Weighted round robin, Connection ID, or Weighted least connections for Load Balancing Algorithm. If you enable sticky sessions, all requests from the same client during one session are sent to the same backend server. For more information about sticky sessions, see Enabling Sticky Session to Accelerate Access. NOTE: TLS backend server groups do not support sticky session. Sticky session is enabled by default and is not shown for QUIC backend server groups.
Sticky Session Type	Specifies the sticky session type. This parameter is mandatory if Sticky Session is enabled. You can select one of the following types: Source IP address: The source IP address of each request is calculated using the consistent hashing algorithm to obtain a unique hashing key, and all backend servers are numbered. The system allocates the client to a particular server based on the generated key. This ensures requests from the same IP address are forwarded to the same backend server. Load balancer cookie: The load balancer generates a cookie after receiving a request from the client. All subsequent requests with the cookie are routed to the same backend server. Application cookie: The application deployed on the backend server generates a cookie after receiving the first request from the client. All subsequent requests with the cookie are routed to the same backend server. NOTE: Source IP address is available when you have selected TCP, QUIC, or UDP for Backend Protocol. Load balancer cookie and Application cookie are available when you have selected HTTP, GRPC, or HTTPS for Backend Protocol.
Stickiness Duration (min)	Specifies the minutes that sticky sessions are maintained. This parameter is mandatory if Sticky Session is enabled. Sticky sessions at Layer 4: 1 to 60 Sticky sessions at Layer 7: 1 to 1440
Slow Start	Specifies whether to enable slow start. This parameter is optional if you have selected Weighted round robin for Load Balancing Algorithm. After you enable this option, the load balancer linearly increases the proportion of requests to backend servers in this mode. When the slow start duration elapses, the load balancer sends full share of requests to backend servers and exits the slow start mode. NOTE: Slow start is only available for HTTP, gRPC, and HTTPS backend server groups of dedicated load balancers. For more information about the slow start, see Slow Start.
Slow Start Duration (s)	Specifies how long the slow start will last, in seconds. This parameter is mandatory if Slow Start is enabled.
Deregistration Delay	This parameter is enabled by default if the backend protocol is TCP, UDP, or QUIC. If a backend server is removed or the health check fails, ELB continues to route in-flight requests to this server until the deregistration delay timeout expires. NOTE: This option is available in certain regions. You can see which regions support this option on the console.
Deregistration Delay Timeout (s)	This parameter is mandatory if Deregistration Delay is enabled. ELB continues to route in-flight requests to the backend server until the deregistration delay timeout expires. The value ranges from 10 to 4000, in seconds. The default value is 300 on the console.
Description (Optional)	Provides supplementary information about the backend server group.

Click Next to add backend servers and configure health check.

Add cloud servers, supplementary network interfaces, or IP as backend servers to this backend server group. For details, see Backend Server Overview.

Configure health check for the backend server group based on Table 4. For more information about health checks, see Health Check.

**Table 4** Parameters required for configuring a health check
Parameter	Description
Health Check	Specifies whether to enable the health check option. If health check is enabled, click next to Advanced Settings (Optional) to set health check parameters.
Health Check Protocol	Specifies the protocol that will be used by the load balancer to check the health of backend servers. TCP, HTTP, TLS, gRPC, and HTTPS are supported. If the protocol of the backend server group is UDP and QUIC, the health check protocol is UDP by default and cannot be changed. NOTE: TLS and gRPC are available in certain regions. You can see which regions support them on the console.
HTTP Method	Specifies the request method for health checks. This parameter is mandatory if the health check protocol is HTTP, HTTPS, or gRPC. GET: The backend server returns all the information. HEAD: The backend server returns only HTTP headers, improving request efficiency. Ensure that your backend servers support HEAD requests. Otherwise, the health check may fail. In this case, you can use GET to perform health checks. POST: Ensure that your backend servers support POST requests. Otherwise, the health check may fail. In this case, you can use GET to perform the health check. NOTE: If the health check protocol is HTTP or HTTPS, the GET and HEAD methods are supported. If the health check protocol is gRPC, the GET and POST methods are supported.
Domain Name	Specifies the domain name that will be used for health checks. This parameter is mandatory if the health check protocol is HTTP, gRPC, or HTTPS. By default, the private IP address of each backend server is used. You can also specify a domain name that consists of at least two labels separated by periods (.). Use only letters, digits, and hyphens (-). Do not start or end strings with a hyphen. Max total: 100 characters. Max label length: 63 characters.
Health Check Port	Specifies the port that will be used by the load balancer to check the health of backend servers. The port number ranges from 1 to 65535. NOTE: By default, the service port on each backend server is used. You can also specify a port for health checks.
Path	Specifies the health check URL, which is the destination on backend servers for health checks. This parameter is mandatory if the health check protocol is HTTP, gRPC, or HTTPS. The path can contain 1 to 80 characters and must start with a slash (/). The path can contain letters, digits, hyphens (-), slashes (/), periods (.), question marks (?), number signs (#), percent signs (%), ampersands (&), and extended character sets _;~!. () *[]@$^:',+
Interval (s)	Specifies the maximum time between two consecutive health checks, in seconds. The interval ranges from 1 to 50.
Timeout (s)	Specifies the maximum time required for waiting for a response from the health check, in seconds. The value ranges from 1 to 50.
Healthy Threshold	Specifies the number of consecutive successful health checks required for declaring a backend server healthy. The value ranges from 1 to 10.
Unhealthy Threshold	Specifies the number of consecutive failed health checks required for declaring a backend server unhealthy. The value ranges from 1 to 10.
Status Code	Specifies the status codes that will be returned to the load balancer to indicate the health of backend servers. This parameter is available only when you set the health check protocol to HTTP, gRPC, or HTTPS. You can enter a unique number or a number range within the status code range, for example, 0-10 and 200-300. A maximum of five HTTP status codes and number ranges are supported. If there is more than one status code or number range, press Enter to separate them. If the health check protocol is HTTP or HTTPS, the status code ranges from 200 to 599. When the gRPC protocol is used, the status code ranges from 0 to 99. NOTE: This feature will be available in more regions. See details on the management console.