What is a Backend Group, and how is it different from a Backend Service?

Short answer

A Backend Group gathers more than one server providing the same service into a single record. Settings such as Health Check, SSL Service, SNAT and Maintenance Mode are defined once at group level and applied as they are to every server in the group; Weight, Connection Limit and Cookie Content, on the other hand, are entered separately per server. A Backend Service defines a single server. Both are used as the target of a vService and are monitored from their own Dashboard screen.

How Does It Work?

• Common settings at group level: Health check, SSL service, SNAT, Proxy, Route table, Backup service and Maintenance mode are defined once in the group; they are automatically applied to all servers in the group. A new server added to the group inherits these settings on its own.
• Individual settings at server level: The address and port, Weight, Connection limit and Cookie content value of each server are entered separately. This way, servers with different capacities are used in the same group.
• Health check: The Health Checks profile selected in the group is run separately for each server in the group; all servers are checked with the same profile (the same check type, Period, Timeout and thresholds), but the result of each server is its own. A failed server alone is taken out of traffic, and the group continues to serve with the remaining healthy servers.
• Group status: The Health indicator of the group is calculated from the results of the servers: if all servers are in the same status, that status is shown (e.g. OK, Error); if the statuses differ from each other, Partially healthy is shown.

Comparison

	Backend Service	Backend Group
Number of servers	A single server	More than one server
Address	A single IP:port	A separate IP:port per server
Common settings	None; each record carries its own setting	Shared at group level
Weight / Connection limit	A single value	A separate value per server
Health check	Its own profile, its own result	A single profile; each server is checked separately, and the group result is calculated collectively (including Partially healthy)
Number of records	Each server is a separate record	All servers in a single record
Monitoring	Backend Service Dashboard	Backend Group Dashboard; in addition, a breakdown per server
When?	When there is a single server, or when the server requires completely different settings	When there is more than one server providing the same service, or when servers are added/removed often

Backend Group Monitoring Screen

In the Left Monitor Panel, when the group is clicked in the Backend Service Groups tab, the Dashboard screen of the group opens:

View tab	What it shows
Overview	The total of the traffic the group members receive from all the vServices in which the group is used (HTTP Requests, Response Time, Bandwidth, Current Sessions, Queue Depth, Connect Time, Connections, HTTP Response Codes)
Health Check	The health check status history and check response time of each group member
Per Backend Service	The breakdown of the same metrics into separate series for each group member; which server slowed down or produced errors is seen here
Per vService	The breakdown of the same metrics into separate series for each vService in which the group is used

Backend Group in TR7

• Adding: It is added with the Load Balancer → Backend Groups → Add steps. Each server in the group is entered as a row in the Backends table in the form.
• Usage: Like a Backend Service, the group is added to the Default Backend Services or Conditional Backend Services list of the vService.
• List view: When the Show Backend Services switch is turned on in the Backend Groups list, the servers of each group appear in separate rows with their own Health status, Weight and address.
• Recommendation: Even if there is only one server now, if servers will be added in the future, creating a group from the start reduces the next addition to just entering an IP:port.

Common Misconceptions

• "A Backend Group gathers existing Backend Services together." The group does not contain Backend Service records; the servers are entered separately into the group's own Backends table. The two are independent objects.
• "The health check of the group checks the group as a single whole." The profile is selected once in the group, but each server is checked separately; only the failed server is taken out of traffic, and the remaining healthy servers share the traffic.
• "If I make the group a backup (Backup service), traffic is distributed across the backup servers." A backup group takes effect only when all main targets are unavailable, and at that moment only the first available backup server receives traffic.
• "I can also put different services in the same group." A group is for servers that provide the same service and can share the same settings; different services such as web and API servers are kept in separate groups.

How it is done in the documentation

• Backend Groups · What Is Backend Groups?
• Backend Groups · Backend Group and Backend Service Comparison
• Backend Groups · Health Status Indicator
• Backend Groups Dashboard · What Is Dashboard?
• Backend Groups · How to Add?

Related questions

• What is a Backend Service?
• When should I use a Backend Group, and when individual Backend Services?
• Should I create a Backend Group even if I only have one server?
• Which settings are shared at group level, and which are per server?