Blog Post

Azure Network Security Blog
6 MIN READ

Azure DDoS Custom Policy GA

Jdasari's avatar
Jdasari
Icon for Microsoft rankMicrosoft
Nov 16, 2026

More control, broader coverage with Azure DDoS Protection custom policy

Fine-tune protocol-specific DDoS detection for internet-facing workloads while keeping Azure’s adaptive protection as the default.

Organizations depend on Azure DDoS Protection to monitor traffic continuously and mitigate network-layer attacks at cloud scale. Most workloads benefit from a hands-off, adaptive experience. But applications with planned traffic surges, unusual packet-rate patterns, strict latency requirements, or distinct environments sometimes need more control over when mitigation begins.

Azure DDoS Protection custom policy adds that optional control. It lets you define fixed inbound detection thresholds for TCP, UDP, and TCP SYN traffic and apply them to supported public endpoints. The result is a protection model that remains automatic by default while giving application and network teams a precise override when their workload calls for it.

In one sentence: Custom policy extends Azure DDoS Protection’s adaptive baseline with optional, protocol-specific controls; it does not replace the turnkey default.

Azure DDoS Protection custom policy is now generally available, letting customers replace machine-learning auto-tuning with fixed protocol-specific detection thresholds per protected resource across Azure resource types with public IPs. Available on Network Protection and IP Protection SKUs, it already covers 929 policies and about 1,970 attachments pre-GA, led by OpenAI. 

Why customers asked for more control

Adaptive auto-tuning works well for most workloads because Azure learns the normal traffic profile of each protected endpoint and adjusts mitigation thresholds as that profile changes. A fixed, application-specific threshold can be useful when traffic is predictable in a way that the service cannot infer ahead of time.

A gaming launch, ticket sale, live stream, or seasonal event is expected to create a sudden but legitimate increase in packet rate.

A high-throughput service regularly operates near a threshold and needs an explicitly controlled trigger.

A latency-sensitive workload benefits from a tighter mitigation threshold for one protocol.

Development, test, and production environments need separate policies and change controls.

Security teams want repeatable, infrastructure-as-code deployment and verifiable associations to protected endpoints.

Automatic by default; explicit where you choose

A custom threshold changes behavior only for the protocol you configure on the associated resource. Azure uses the fixed value for that protocol, while protocols without a custom rule continue to use adaptive auto-tuning. Detaching the custom policy returns the endpoint to the default adaptive behavior.

Design principle: Start with the default adaptive service. Add a custom rule only when you have a measured workload requirement and a monitoring plan.

 

What is new in the expanded experience

The launch experience is designed around the Azure resource and its public endpoint, rather than asking customers to reason about an internal “frontend IP” abstraction. The underlying association still depends on the resource and management interface, but the customer journey begins with the workload that needs protection.

Broader resource reach. Apply a custom policy to supported public endpoints across virtual machines, virtual machine scale sets, Azure Kubernetes Service, Standard Load Balancer, Application Gateway, Azure Firewall, and Azure Bastion.

A public IP-centric portal flow. Choose eligible public IP addresses from one picker, regardless of the supported resource behind them.

Consistent management options. Create and manage policies through the Azure portal, Azure CLI, Azure PowerShell, and ARM/REST, subject to management-surface rollout.

Clear protocol controls. Set independent inbound detection thresholds for TCP, UDP, and TCP SYN traffic.

Existing observability. Continue using Azure Monitor metrics, diagnostic logs, alerts, and DDoS Protection telemetry to validate behavior and investigate mitigation activity.

Resource-oriented association model

Resource

Customer-facing scope

Underlying association

Azure Virtual Machines

VM public endpoint

Instance-level public IP

Virtual Machine Scale Sets

VMSS public endpoint

Individual public IP, subject to ownership and lifecycle

Azure Kubernetes Service

AKS public endpoint

Public IP or Standard Load Balancer frontend

Standard Load Balancer

Public frontend

Public IP or frontend IP configuration, depending on management surface

Application Gateway

Application public endpoint

Associated public IP configuration

Azure Firewall

Firewall public endpoint

Associated public IP configuration

Azure Bastion

Bastion public endpoint

Associated public IP configuration

 

How detection threshold rules work

Each rule names a traffic type and the inbound packet-per-second (PPS) value that triggers mitigation. The currently validated ranges are:

Traffic type

Valid range

Example use

TCP

50,000–2,000,000 PPS

Control general inbound TCP trigger

UDP

20,000–200,000 PPS

Tune for packet-intensive UDP services

TCP SYN

1,000–100,000 PPS

Tune connection-establishment trigger

Treat the range as a validation boundary, not a recommendation. Select a value from measured normal traffic, expected peak traffic, application capacity, and your risk tolerance.

Create a policy in the Azure portal

The portal workflow brings policy creation, public IP selection, rules, tags, and review into a single guided experience.

In the Azure portal, search for DDoS custom policies and select Create.

Choose the subscription, resource group, policy name, and region. Select the eligible public IP addresses that should use the policy.

 

 

Figure 1. Create the policy and select public IP addresses. Demo values are sanitized.

Use the public IP picker to select supported endpoints. The portal resolves the associated resource type and shows existing custom-policy associations.

 

 

 

 

Figure 2. Choose eligible public IP addresses across supported resources. Resource values are redacted.

On Custom rules, add the protocol-specific thresholds your workload requires. You do not have to define all protocols.

 

 

Figure 3. Configure independent inbound TCP, UDP, and TCP SYN thresholds.

Review the selected endpoints and rule count, then create the policy.

 

 

Figure 4. Review the policy configuration before deployment. Demo values are sanitized.

Automate with Azure CLI

The following example creates a custom policy with an inbound TCP threshold of 1,000,000 PPS and then returns the provisioning state and detection rules.

Create one detection rule

az network ddos-custom-policy create \
  --resource-group MyResourceGroup \
  --name MyCustomPolicy \
  --location eastus \
  --detection-rule-name tcp-threshold \
  --detection-mode TrafficThreshold \
  --traffic-type Tcp \
  --packets-per-second 1000000

 

Verify the policy

az network ddos-custom-policy show \
  --resource-group MyResourceGroup \
  --name MyCustomPolicy \
  --query '{name:name, state:provisioningState, rules:detectionRules[].trafficDetectionRule}'

 

Association commands differ by resource and management surface. In the validated launch guidance, Azure CLI and Azure PowerShell are two management paths to the same capability: CLI can associate the policy through the eligible public IP, while PowerShell can use the resource-specific frontend configuration for Load Balancer. Follow the current Learn procedure for the resource you are configuring.

Automate with Azure PowerShell

Azure PowerShell can create several protocol rules in one policy.

Create a policy with TCP, UDP, and TCP SYN rules

$tcpRule = New-AzDdosCustomPolicyDetectionRule `
  -Name 'tcpRule1' -TrafficType Tcp -PacketsPerSecond 1000000
$udpRule = New-AzDdosCustomPolicyDetectionRule `
  -Name 'udpRule1' -TrafficType Udp -PacketsPerSecond 100000
$tcpSynRule = New-AzDdosCustomPolicyDetectionRule `
  -Name 'tcpSynRule1' -TrafficType TcpSyn -PacketsPerSecond 50000

$policy = New-AzDdosCustomPolicy `
  -ResourceGroupName 'MyResourceGroup' `
  -Name 'MyCustomPolicy' `
  -Location 'eastus' `
  -DetectionRule @($tcpRule, $udpRule, $tcpSynRule)

 

 

Operational guidance

Measure before you set. Use normal and peak traffic data, not a generic threshold copied from another application.

Leave headroom. Account for organic growth, failover, scale-out, and known events.

Change one protocol at a time. This makes results easier to observe and rollback.

Test outside production first. Validate policy creation, association, monitoring, and detachment in a lower environment.

Monitor after every change. Watch DDoS metrics, mitigation notifications, flow logs, and application health.

Document ownership. Treat threshold and mitigation-rule changes as production security changes with RBAC, review, and change history.

Know the rollback. Detach the policy or remove a protocol rule to return that traffic to adaptive auto-tuning, following the current product procedure.

Get started

Start with the overview, choose the management surface that fits your workflow, deploy a test policy, and validate telemetry before production rollout.

Azure DDoS Protection custom policy overview

Create and manage a custom policy in the Azure portal

Create a custom policy using Azure CLI

Azure CLI command reference: az network ddos-custom-policy

Azure PowerShell command reference: New-AzDdosCustomPolicy

Monitor Azure DDoS Protection

Azure DDoS Protection remains designed to be automatic and adaptive at scale. Custom policy adds targeted control for the workloads that need it—without turning every protected application into a manually tuned system.

Updated Nov 16, 2026
Version 1.0
No CommentsBe the first to comment