Resource Dependencies and Ordering
What You'll Learn
You'll learn how ARM templates determine which Azure resources to deploy first and how to explicitly control that sequence when needed.
Understanding Deployment Order
When you declare multiple resources in an ARM template, Azure needs to know which ones to create first. Some resources naturally depend on others—you can't configure a VM's network interface before the virtual network exists, for example.
ARM templates handle dependencies in two ways:
Implicit Dependencies
These happen automatically when you reference another resource's properties using functions like reference() or resourceId(). Azure detects these references and creates resources in the correct order without you specifying anything extra.
Analogy: It's like assembling furniture—when instructions say "attach piece B to piece A," you automatically know to build piece A first.
Explicit Dependencies with dependsOn
Sometimes the relationship isn't obvious from property references alone. The dependsOn array lets you declare explicit dependencies:
{
"type": "Microsoft.Compute/virtualMachines",
"apiVersion": "2021-03-01",
"name": "myVM",
"dependsOn": [
"[resourceId('Microsoft.Network/virtualNetworks', 'myVNet')]"
]
}
Use dependsOn when:
- Resource A must exist before B, but B doesn't reference A's properties
- You need to enforce ordering for operational reasons (like ensuring a database initializes before an app server)
- You're using deployment scripts that need specific sequencing
Azure creates resources in parallel when possible, deploying dependent resources only after their prerequisites complete.
Key Takeaway: ARM templates use implicit dependencies (detected from property references) and explicit dependsOn declarations to ensure resources deploy in the correct order, enabling safe parallel deployment while respecting required sequences.