Azure Load Balancer components
Azure Load Balancer includes a few key components. These components can be configured in your subscription through the Azure portal, Azure CLI, Azure PowerShell, Resource Manager Templates or appropriate alternatives.
The IP address of your Azure Load Balancer. It's the point of contact for clients. These IP addresses can be either:
- Public IP Address
- Private IP Address
The nature of the IP address determines the type of load balancer created. Private IP address selection creates an internal load balancer. Public IP address selection creates a public load balancer.
|Public load balancer||Internal load balancer|
|Frontend IP configuration||Public IP address||Private IP address|
|Description||A public load balancer maps the public IP and port of incoming traffic to the private IP and port of the VM. Load balancer maps traffic the other way around for the response traffic from the VM. You can distribute specific types of traffic across multiple VMs or services by applying load-balancing rules. For example, you can spread the load of web request traffic across multiple web servers.||An internal load balancer distributes traffic to resources that are inside a virtual network. Azure restricts access to the frontend IP addresses of a virtual network that are load balanced. Front-end IP addresses and virtual networks are never directly exposed to an internet endpoint, meaning an internal load balancer cannot accept incoming traffic from the internet. Internal line-of-business applications run in Azure and are accessed from within Azure or from on-premises resources.|
|SKUs supported||Basic, Standard||Basic, Standard|
Load balancer can have multiple frontend IPs. Learn more about multiple frontends.
The group of virtual machines or instances in a virtual machine scale set that is serving the incoming request. To scale cost-effectively to meet high volumes of incoming traffic, computing guidelines generally recommend adding more instances to the backend pool.
Load balancer instantly reconfigures itself via automatic reconfiguration when you scale instances up or down. Adding or removing VMs from the backend pool reconfigures the load balancer without additional operations. The scope of the backend pool is any virtual machine in a single virtual network.
Backend pools support addition of instances via network interface or IP addresses.
When considering how to design your backend pool, design for the least number of individual backend pool resources to optimize the length of management operations. There's no difference in data plane performance or scale.
A health probe is used to determine the health status of the instances in the backend pool. During load balancer creation, configure a health probe for the load balancer to use. This health probe will determine if an instance is healthy and can receive traffic.
You can define the unhealthy threshold for your health probes. When a probe fails to respond, the load balancer stops sending new connections to the unhealthy instances. A probe failure doesn't affect existing connections. The connection continues until the application:
- Ends the flow
- Idle timeout occurs
- The VM shuts down
Load balancer provides different health probe types for endpoints: TCP, HTTP, and HTTPS. Learn more about Load Balancer Health probes.
Basic load balancer doesn't support HTTPS probes. Basic load balancer closes all TCP connections (including established connections).
Load Balancer rules
A load balancer rule is used to define how incoming traffic is distributed to all the instances within the backend pool. A load-balancing rule maps a given frontend IP configuration and port to multiple backend IP addresses and ports. Load Balancer rules are for inbound traffic only.
For example, use a load balancer rule for port 80 to route traffic from your frontend IP to port 80 of your backend instances.
Figure: Load-Balancing rules
High Availability Ports
A load balancer rule configured with 'protocol - all and port - 0' is known as an High Availability (HA) port rule. This rule enables a single rule to load-balance all TCP and UDP flows that arrive on all ports of an internal Standard Load Balancer.
The load-balancing decision is made per flow. This action is based on the following five-tuple connection:
- source IP address
- source port
- destination IP address
- destination port
The HA ports load-balancing rules help you with critical scenarios, such as high availability and scale for network virtual appliances (NVAs) inside virtual networks. The feature can help when a large number of ports must be load-balanced.
Figure: HA Ports rules
Learn more about HA ports.
Inbound NAT rules
An inbound NAT rule forwards incoming traffic sent to frontend IP address and port combination. The traffic is sent to a specific virtual machine or instance in the backend pool. Port forwarding is done by the same hash-based distribution as load balancing.
Figure: Inbound NAT rules
Inbound NAT rules in the context of Virtual Machine Scale Sets are inbound NAT pools. Learn more about Load Balancer components and virtual machine scale set.
An outbound rule configures outbound Network Address Translation (NAT) for all virtual machines or instances identified by the backend pool. This rule enables instances in the backend to communicate (outbound) to the internet or other endpoints.
Learn more about outbound connections and rules.
Basic load balancer doesn't support outbound rules.
Figure: Outbound rules
- Learn about load balancer limits
- Load balancer provides load balancing and port forwarding for specific TCP or UDP protocols. Load-balancing rules and inbound NAT rules support TCP and UDP, but not other IP protocols including ICMP.
- Load Balancer backend pool cannot consist of a Private Endpoint.
- Outbound flow from a backend VM to a frontend of an internal Load Balancer will fail.
- A load balancer rule cannot span two virtual networks. All load balancer frontends and their backend instances must be in a single virtual network.
- Forwarding IP fragments isn't supported on load-balancing rules. IP fragmentation of UDP and TCP packets isn't supported on load-balancing rules. HA ports load-balancing rules can be used to forward existing IP fragments. For more information, see High availability ports overview.
- You can only have 1 Public Load Balancer and 1 internal Load Balancer per availability set
- See Create a public Standard load balancer to get started with using a Load Balancer.
- Learn more about Azure Load Balancer.
- Learn about Public IP Address
- Learn about Private IP Address
- Learn about using Standard load balancer and Availability Zones.
- Learn about Standard load balancer Diagnostics.
- Learn about TCP Reset on Idle.
- Learn about Standard load balancer with HA Ports load balancing rules.
- Learn more about Network Security Groups.
- Learn more about Load balancer limits.
- Learn about using Port forwarding.
Submit and view feedback for