Azure compute resources that are used to create and manage groups of heterogeneous load-balanced virtual machines.
Yes — the important point is that the single Public IP shown in the VMSS Overview is not a reliable indication that the VMSS has only one public entry point.
In your topology:
Standard Load Balancer
┌──────────────────────────┐
│ Frontend IP 1 → Public IP A
Internet ─┤ Frontend IP 2 → Public IP B
│ Frontend IP 3 → Public IP C
└────────────┬─────────────┘
│
Load-balancing rules
│
Same Backend Pool
│
┌────────┴────────┐
│ VMSS │
│ VM1 VM2 VM3 ... │
└─────────────────┘
Azure Load Balancer explicitly supports multiple frontend IP configurations, and each frontend is an independent entry point. A load-balancing rule associates a particular frontend IP/port with a backend pool. Microsoft Learn
So why does VMSS Overview show only one IP?
The VMSS Overview blade's Public IP field is a portal/UI summary value, rather than a representation of every frontend IP associated with the Load Balancer.
I would not interpret the displayed IP as “the VMSS's active public IP” or as Azure selecting that IP for load balancing.
The actual networking configuration is determined by the Load Balancer's:
- Frontend IP configurations
- Load-balancing rules
- Backend pools
- Health probes
- Outbound rules, where applicable
Microsoft's documentation describes the frontend IP as the Load Balancer's point of contact and confirms that a Load Balancer can have multiple frontend IPs. Microsoft Learn
Is there a documented selection algorithm?
I would not rely on or document a deterministic Azure selection rule for the VMSS Overview IP.
In particular, don't assume:
- first frontend IP = displayed IP
- primary frontend = displayed IP
- lowest numerical IP = displayed IP
- first load-balancing rule = displayed IP
- default frontend = displayed IP
Those are not documented as the networking behavior you should depend upon.
Instead, consider the Overview value as portal display/summary information. For the authoritative configuration, inspect the Load Balancer itself.
How to verify the real configuration
In Azure Portal:
Load Balancer → Frontend IP configurations
You should see something like:
Frontend IP configurations
-----------------------------------------
Frontend-IP-1 20.x.x.x
Frontend-IP-2 40.x.x.x
Frontend-IP-3 52.x.x.x
Then:
Load Balancer → Load balancing rules
For example:
Rule Frontend Backend Pool
------------------------------------------------
HTTP-Rule Public-IP-A VMSS-Pool
HTTPS-Rule Public-IP-B VMSS-Pool
TCP-Rule Public-IP-C VMSS-Pool
That is what determines which public IP accepts which traffic — not the IP displayed on the VMSS Overview blade.
Microsoft's multiple-frontend documentation specifically describes this model: each frontend is referenced by rules, and a load-balancing rule distributes traffic arriving on that frontend across healthy backend instances. Microsoft Learn
Bottom line: If the VMSS Overview shows one Public IP while the Standard Load Balancer has multiple frontend Public IPs, treat the Overview IP as a portal summary value with no special traffic-selection significance. The Load Balancer frontend/rule configuration is authoritative. Microsoft Learn
Microsoft Learn — Multiple frontends for Azure Load Balancer
Microsoft Learn — Manage a public IP address with a load balancer