Application Gateway v2 Availability Zone Pricing

Question

Application Gateway v2 Availability Zone Pricing

FU JIE ZHAO 120

Hi Support，

I have a few questions about this document:

https://learn.microsoft.com/en-us/azure/reliability/reliability-application-gateway-v2

My current dev environment aag is v1, with only one instance and no Availability Zone enabled.

I'm now preparing to migrate to standard_v2, but when I use the cloning script, Availability Zones 1, 2, and 3 are enabled by default.

I noticed the documentation mentions, "All gateways have a minimum of two instances for high availability. Even if the Azure portal indicates that your gateway has a single instance, internally it's always created with a minimum of two instances."

User's image

Does this mean that enabling Availability Zones by default uses two instances? Even if I adjust it to one instance now, will I still be charged for two instances?

Price for 1 instance: User's image

Prices for 2 instances:

User's image

One instance corresponds to 10 Compute units(s), right?

0 comments

Answer accepted by question author

1 additional answer

Your answer

Answer 1

Ravi Varma Mudduluru 9,200 Microsoft External Staff Moderator

Hello @ FU JIE ZHAO,

Thanks for reaching out to Microsoft Q&A.

The cloning script is working exactly as designed. In any region that supports Availability Zones, new Standard_v2 gateways are created zone-redundant by default across all available zones (1, 2, and 3 in your case). This is Microsoft’s recommended high-availability setup and not something you need to change unless you specifically want a single-zone gateway for dev cost-matching purposes.

On your two main questions:

Yes, every v2 gateway always runs with a minimum of two instances internally for high availability, even if the portal shows “1”. Enabling zones simply spreads those instances across the selected zones — it doesn’t add extra instances or extra cost.
Billing is the same whether zones are enabled or not. The Fixed Gateway Hours charge already covers the built-in HA (the two internal instances). If you use Autoscaling and set the minimum instance count to 0, you pay only the fixed cost when idle. Each visible “instance” in the scale settings equals 10 Capacity Units, which matches what you’re seeing in the pricing calculator.

Recommended solution for your dev environment (lowest cost + full HA):

Keep the cloned zone-redundant gateway.
Go to Scale settings → choose Autoscaling.
Set Minimum instance count = 0 (or 1 if you want a small warm buffer).
Save.

This gives you the modern v2 benefits (better performance, future WAF_v2 option, etc.) while keeping idle costs minimal.

If you really want to match your old v1 single-zone setup exactly, you’ll need to delete the cloned gateway and create a new one manually (zones can’t be changed after creation), selecting only Zone 1 under Availability zones. But for most dev workloads, the default zone-redundant config is the way to go.

Here are the exact Microsoft docs that cover this:

If the answer is helpful, please click "Accept Answer" and kindly upvote it. If you have extra questions about this answer, please click "Comment".

FU JIE ZHAO 120 Reputation points

2026-03-26T06:13:52.59+00:00

Thank you @Ravi Varma Mudduluru .

You mean that with Availability Zones enabled by default, I changed from 2 instances to 1, but internally two instances are actually running. However, I only need to be responsible for the cost of 1 instance? Besides the fixed costs, for variable costs, I only need to be responsible for:

Region: East Asia

Variable Costs = $0.008 * 1 (Instance count) * 10 (capacity units) * 730 (Hours) = $58.4/month

instead of 20 (capacity units)?
FU JIE ZHAO 120 Reputation points

2026-03-26T06:23:54.9466667+00:00

I have already set the number of AAG instances to 1 on the portal, but I am currently unable to determine the exact charges.
Ravi Varma Mudduluru 9,200 Reputation points Microsoft External Staff Moderator

2026-03-26T06:45:18.9866667+00:00
Hello @ FU JIE ZHAO,

Thanks for the quick follow-up.

To clarify directly:

No, you do not get charged for two instances just because Availability Zones are enabled and the platform guarantees internal high availability.

Here's how billing actually works for Standard_v2:

For the variable Capacity Unit cost, you are only billed based on the Minimum instance count you set in the Autoscaling configuration.

Each configured instance reserves 10 Capacity Units.

So if you set Minimum instance count = 1, you pay for 10 CU (your calculation of ~$58.40/month at $0.008 per CU-hour is correct).

You do not pay for 20 CU (two instances) unless you explicitly set the minimum to 2 or higher.

Best practice for your dev environment:

Set the scaling mode to Autoscaling and Minimum instance count = 0. This removes almost all variable (capacity) costs when the gateway is idle, while the platform still provides the built-in high availability (covered in the fixed cost) and keeps the zone-redundant protection across zones 1, 2, and 3.

You can always increase the minimum later if you need a small warm buffer.

If the answer is helpful, please click "Accept Answer" and kindly upvote it. If you have extra questions about this answer, please click "Comment".
FU JIE ZHAO 120 Reputation points

2026-03-26T06:59:43.13+00:00

Hi @Ravi Varma Mudduluru ,

It seems I misunderstood the instructions in the documentation. Does this mean that, regardless of whether an availability zone is configured, the fixed fee already covers the cost of two instances? Is that correct?
Ravi Varma Mudduluru 9,200 Reputation points Microsoft External Staff Moderator

2026-03-26T07:09:54.15+00:00

Hello @ FU JIE ZHAO

Yes, that is correct.

For Azure Application Gateway Standard_v2 (and WAF_v2), the Fixed Gateway Hours charge already covers the cost of the built-in high availability, which internally runs with a minimum of two instances regardless of whether Availability Zones (zone redundancy) are configured or not.

The fixed cost remains the same whether zones are enabled or not. You only pay extra for actual usage through Capacity Units when traffic increases.

If the answer is helpful, please click "Accept Answer" and kindly upvote it. If you have extra questions about this answer, please click "Comment".

Answer 2

Availability Zones and instance count are independent of pricing for Application Gateway v2. Pricing is based on:

A fixed hourly cost for the gateway (per gateway, per region), independent of:
- How many instances are running
- How many Availability Zones are used (zonal or zone‑redundant)
Capacity unit consumption or reservation.

From the v2 pricing model:

Fixed costs are charged as long as the gateway is provisioned, “regardless of the number of instances running within the same Azure region.”
Capacity unit costs are based on the number of capacity units reserved or actually used.
When instance count is set manually or via autoscale minimum, each instance reserves 10 capacity units. These reserved capacity units are billed as long as the gateway is active, even if actual usage is lower.

How this applies to the questions:

“All gateways have a minimum of two instances” vs. billing
The statement about “a minimum of two instances” refers to the internal high‑availability implementation of the service. It doesn’t mean two billable instances are exposed or that zone‑redundant configuration automatically doubles the fixed gateway charge. The fixed v2 gateway cost is per gateway, not per instance.
Effect of enabling Availability Zones (1, 2, 3)
Enabling zone‑redundant deployment (zones 1, 2, 3) does not change the fixed gateway price. Availability zone support “doesn't incur extra charges beyond the standard capacity unit pricing.” The service may run instances across multiple zones for resiliency, but the fixed cost remains the same.
“If I set instance count to 1, am I charged for 2?”
The billing model does not charge per instance; it charges:
- One fixed gateway fee per hour, plus
- Capacity units. If instance count is set to 1, that reserves 10 capacity units. Billing is for those 10 reserved capacity units (and any additional capacity units if actual usage exceeds them), not for “two instances.”
“One instance corresponds to 10 compute units, right?”
Yes. For v2 SKUs, “each instance guarantees a minimum of 10 capacity units in terms of processing capability.” When an instance is reserved (via manual scale or autoscale minimum), 10 capacity units are reserved and billed for that instance.

In summary: enabling Availability Zones does not add extra instance‑based cost. Costs are fixed gateway fee + capacity units, and each configured instance reserves 10 capacity units.

References:

Share via

Application Gateway v2 Availability Zone Pricing

1 additional answer

Your answer