Cloud Armor on Google Cloud: Edge Security Policies and the CDN Use Case

GCP Study Hub
November 12, 2026

Cloud Armor protects HTTP and HTTPS applications that are fronted by Google Cloud's external HTTPS load balancers. It enforces security policies that safeguard applications and external services from web-based threats, and because those policies are applied at the load balancer, attacks are mitigated before they reach backend services. For the Professional Cloud Architect exam, Cloud Armor is relevant to scenarios about securing public-facing applications, and to questions about how it differs from Cloud Firewall rules.

What Cloud Armor provides

  • Protection against distributed denial of service (DDoS) attacks, which helps applications remain available even under high traffic spikes.
  • IP filtering, which allows or denies traffic based on the source IP address range.
  • Geo-based access restrictions, which control which users or regions can access a service.

Cloud Armor security policies integrate directly with Google Cloud's load balancing infrastructure, which makes Cloud Armor useful for protecting applications at the network edge while keeping access control flexible. Cloud Load Balancing can front Compute Engine, Google Kubernetes Engine, and serverless services such as Cloud Run and App Engine (sometimes through a network endpoint group), so a Cloud Armor policy on the load balancer can protect those backends. With a load balancer placed in front of a Cloud Run service, for example, Cloud Armor can be configured on that load balancer to block certain IP addresses, enforce rate limiting, and prevent DDoS attacks.

In case study architectures, Cloud Armor typically fills the requirement for protection against DDoS and web attacks, while Cloud IAM handles who is allowed to access what.

Managing Cloud Armor with gcloud

Cloud Armor is part of the broader Compute Engine ecosystem, so there is no standalone gcloud armor command. Cloud Armor configurations are managed through security policies, in the same way that there is no direct gcloud vpc or gcloud network command because networking also lives under Compute Engine. Cloud Armor commands start with gcloud compute security-policies.

A rule that allows traffic from a specific IP range is created with gcloud compute security-policies rules create, followed by a priority level and an expression that defines the IP range. The rule specifies an action, such as allow, so that traffic from that range is permitted, and it can include a description for clarity and reference. The exact command does not need to be memorized. It is more useful to understand how Cloud Armor rules are structured in gcloud, and that supplying an IP address range lets a rule either allow or deny traffic from that range before it reaches backend services. Besides ranges supplied directly in a rule, a security policy can also use a preconfigured expression to evaluate requests against preconfigured IP lists.

Cloud Armor vs Cloud Firewall rules

Cloud Armor and Cloud Firewall rules both control access to resources, so they can look similar at first. They serve different purposes and operate at different levels.

  • Cloud Firewall rules operate at the network level. They control access to VMs and subnets by filtering traffic based on IP addresses, protocols, and ports. They apply to incoming and outgoing traffic at the infrastructure level, allowing or denying connections before they reach a workload.
  • Cloud Armor works at the application level. It is designed specifically for HTTP and HTTPS traffic passing through external HTTPS load balancers. Instead of managing network-layer access, it enforces security policies for web applications, protecting them from threats such as DDoS attacks and unauthorized access, using IP filtering and geo-based restrictions.

Firewall rules are the broader tool for securing network resources, while Cloud Armor provides targeted protection for web applications. In a Professional Cloud Architect design scenario, the kind of traffic is the guide. Control over which IPs, protocols, and ports can reach VMs and subnets points to firewall rules, and protection for a web application behind an external HTTPS load balancer points to Cloud Armor.

Restricting backends to trusted sources

IP filtering also supports a broader pattern, which is letting only a known set of upstream sources reach an application while every other source is turned away. A content delivery network in front of an application is a typical case. The pattern is not specific to Cloud CDN, and it applies to any upstream provider whose traffic arrives from known IP ranges.

The policy works as an allow list. Cloud Armor evaluates each incoming request against the approved IP ranges, and the contents of that list determine which traffic gets through. Requests that match pass through the external load balancer and reach the backend services in the VPC. Requests from any other source are blocked at the load balancer and never reach the VPC.

Because enforcement happens at the load balancer, the backends only receive traffic from sources the policy trusts, and potentially harmful or unnecessary requests are filtered out before they arrive. The same idea applies whenever a design calls for an application to accept traffic only from specific, known origins.

Our Professional Cloud Architect course covers Cloud Armor alongside Cloud Firewall rules and Cloud Load Balancing, with practice questions that drill these distinctions.

Get tips and updates from GCP Study Hub

arrow