azurerm_service_plan cost estimation

An azurerm_service_plan is the modern resource for an App Service plan, the compute that Web Apps, API Apps, and Function Apps run on. (It replaced the older azurerm_app_service_plan.) Cost is per instance-hour at the plan's tier rate × the worker (instance) count: a Basic B1 plan with one worker is ~$0.017/hour, ~$12.41/month; Standard S1 is ~$70/month; Premium P1v3 ~$120+/month.

The defining cost fact, as with any App Service plan, is that you pay for the plan, not the apps, multiple Web Apps and Function Apps can share one plan at no additional compute charge. So the common overspend is a separate plan per app, multiplying the per-instance fee for capacity that sits mostly idle. Consolidating low-traffic apps onto a shared plan is the biggest lever.

The other decisions are tier (pick the lowest with the features you need, Free/Shared for experiments, Basic for dev, Standard for autoscale/slots, Premium v3 for performance/VNet) and instance count (use autoscale rather than a fixed peak count). Premium v3 is reservable for steady production.

c3x prices the plan from sku_name and worker_count, so the per-instance cost and the effect of scaling out are visible before deployment.

Terraform example

A minimal but realistic configuration that C3X can estimate.

resource "azurerm_service_plan" "main" {
  name                = "app-plan"
  resource_group_name = azurerm_resource_group.main.name
  location            = azurerm_resource_group.main.location
  os_type             = "Linux"
  sku_name            = "B1"
  worker_count        = 1
}

Pricing dimensions

What you actually pay for when you provision azurerm_service_plan.

Dimension	Unit	What's being charged
App Service plan	per instance-hour	Tier rate × worker (instance) count, billed continuously. Apps on the plan share its compute at no extra charge. $0.017/hour for B1 → 1 worker ≈ $12.41/month; S1 ~$70, P1v3 ~$120+

Sample C3X output

A Basic B1 plan with one worker:

azurerm_service_plan.main
└─ App Service plan (B1 × 1)   730 hours   $12.41
                               Monthly     $12.41

Optimization tips

Common ways to reduce azurerm_service_plan cost without changing the workload.

Share one plan across multiple apps

Per avoided plan

You pay for the plan, not the apps, many Web Apps and Function Apps share a plan's compute for free. The common overspend is a plan per app; consolidating low-traffic apps onto a shared plan is the biggest lever.

Right-size the tier

Difference between tiers

Pick the lowest tier with the feature you need: Basic for dev/custom domains, Standard for autoscale and staging slots, Premium v3 for performance and VNet. Paying Premium for a Standard workload is pure overspend.

Autoscale instead of a fixed worker count

Worker-hours during low load

Standard and Premium support autoscale. A plan pinned at peak worker_count pays for peak 24/7; autoscaling to demand pays for peak only when it happens.

Reserve Premium v3 and stop non-prod off-hours

30 to 55% reserved; ~60% off-hours on non-prod

Reserve steady Premium v3 plans (1-3 year) for 30-55% off, and stop or scale down dev/test plans nights and weekends.

FAQ

How is an Azure App Service Plan billed?

Per instance-hour at the tier rate × worker count, continuously, B1 with one worker is ~$12.41/month, S1 ~$70, P1v3 ~$120+. You pay for the plan; multiple apps share it at no extra compute charge.

Is azurerm_service_plan different from azurerm_app_service_plan?

It's the modern replacement, azurerm_service_plan supersedes the older azurerm_app_service_plan. The pricing model is the same: per-instance by tier, with apps sharing the plan.

How does c3x estimate the cost?

From sku_name and worker_count, pricing per-instance hours at the tier rate. Apps on the plan don't add compute cost; scaling out workers (or autoscaling) does.

Related resources

Azure

azurerm_linux_function_app

Function Apps run on a Service Plan (or Consumption)