Skip to main content

NetworkPolicy automation

Otterize automates pod-to-pod access control with network policies, within your cluster.

Instead of managing pod identities, labeling clients, servers and namespaces, and manually authoring individual network policies, Otterize implements intent-based access control (IBAC). You just declare what calls the client pods intend to make, and everything is automatically wired together so only intended calls are allowed.

In this tutorial, we will:

  • Deploy a server pod, and two client pods calling it.
  • Declare that the first client intends to call the server.
  • See that a network policy was autogenerated to allow just that, and block the (undeclared) calls from the other client.


Prepare a Kubernetes cluster that supports network policies

Before you start, you'll need a Kubernetes cluster with a CNI that supports NetworkPolicies.

Below are instructions for setting up a Kubernetes cluster with network policies. If you don't have a cluster already, we recommend starting out with a Minikube cluster.

If you don't have the Minikube CLI, first install it.

Then start your Minikube cluster with Calico, in order to enforce network policies.

minikube start --cpus=4 --memory 4096 --disk-size 32g --cni=calico

The increased CPU, memory and disk resource allocations are required to be able to deploy the ecommerce app used in the visual tutorials successfully.

You can now install (or reinstall) Otterize in your cluster, and optionally connect to Otterize Cloud. Connecting to Cloud lets you:

  1. See what's happening visually in your browser, through the "access graph";
  2. Avoid using SPIRE (which can be installed with Otterize) for issuing certificates, as Otterize Cloud provides a certificate service.

So either forego browser visualization and:

Install Otterize in your cluster, without Otterize Cloud

You'll need Helm installed on your machine to install Otterize as follows:

helm repo add otterize
helm repo update
helm install otterize otterize/otterize-kubernetes -n otterize-system --create-namespace

This chart is a bundle of the Otterize intents operator, Otterize credentials operator, Otterize network mapper, and SPIRE. Initial deployment may take a couple of minutes. You can add the --wait flag for Helm to wait for deployment to complete and all pods to be Ready, or manually watch for all pods to be Ready using kubectl get pods -n otterize-system -w.

After all the pods are ready you should see the following (or similar) in your terminal when you run kubectl get pods -n otterize-system:

NAME                                                       READY  STATUS  RESTARTS AGE
credentials-operator-controller-manager-6c56fcfcfb-vg6m9 2/2 Running 0 9s
intents-operator-controller-manager-65bb6d4b88-bp9pf 2/2 Running 0 9s
otterize-network-mapper-779fffd959-twjqd 1/1 Running 0 9s
otterize-network-sniffer-65mjt 1/1 Running 0 9s
otterize-spire-agent-lcbq2 1/1 Running 0 9s
otterize-spire-server-0 2/2 Running 0 9s
otterize-watcher-b9bf87bcd-276nt 1/1 Running 0 9s

Or choose to include browser visualization and:

Install Otterize in your cluster, with Otterize Cloud

Create an Otterize Cloud account

If you don't already have an account, browse to to set one up.

If someone in your team has already created an org in Otterize Cloud, and invited you (using your email address), you may see an invitation to accept.

Otherwise, you'll create a new org, which you can later rename, and invite your teammates to join you there.

Install Otterize OSS, connected to Otterize Cloud

If no Kubernetes clusters are connected to your account, click the "connect your cluster" button to:

  1. Create a Cloud cluster object, specifying its name and the name of an environment to which all namespaces in that cluster will belong, by default.
  2. Connect it with your actual Kubernetes cluster, by clicking on the "Connection guide " link and running the Helm commands shown there.
    1. Follow the instructions to install Otterize with enforcement on (use the toggle to make Enforcement mode: active)
More details, if you're curious

Connecting your cluster simply entails installing Otterize OSS via Helm, using credentials from your account so Otterize OSS can report information needed to visualize the cluster.

The credentials will already be inlined into the Helm command shown in the Cloud UI, so you just need to copy that line and run it from your shell. If you don't give it the Cloud credentials, Otterize OSS will run fully standalone in your cluster you just won't have the visualization in Otterize Cloud.

The Helm command shown in the Cloud UI also includes flags to turn off enforcement: Otterize OSS will be running in "shadow mode," meaning that it will show you what would happen if it were to create/update your access controls (Kubernetes network policies, Kafka ACLs, Istio authorization policies, etc.). While that's useful for gradually rolling out IBAC, for this tutorial we go straight to active enforcement.

Deploy the server and the two clients

Our simple example consists of three pods: an HTTP server and two clients that call it.

Expand to see the example YAML files
apiVersion: v1
kind: Namespace
name: otterize-tutorial-npol
  1. Deploy the two clients and the server in their namespace using kubectl:
kubectl apply -f
Optional: check deployment status
Check that the client and server pods were deployed
kubectl get pods -n otterize-tutorial-npol

You should see

NAME                           READY   STATUS    RESTARTS   AGE
client-596bcb48d5-pnjxc 1/1 Running 0 8s
client-other-f56d65d7f-z2wg2 1/1 Running 0 8s
server-6bb4784ccc-wtz7f 1/1 Running 0 8s

Let's monitor both client attempts to call the server with additional terminal windows, so we can see the effects of our changes in real time.

  1. Open a new terminal window [client] and tail the client log:
kubectl logs -f --tail 1 -n otterize-tutorial-npol deploy/client
Expected output

At this point the client should be able to communicate with the server:

Calling server...
HTTP/1.1 200 OK
Hi, I am the server, you called, may I help you?
  1. Open another terminal window [client-other] and tail the client-other log:
kubectl logs -f --tail 1 -n otterize-tutorial-npol deploy/client-other
Expected output

At this point the client should be able to communicate with the server:

Calling server...
HTTP/1.1 200 OK
Hi, I am the server, you called, may I help you?

If you've attached Otterize OSS to Otterize Cloud, you can now browse to your account at and see the access graph for your cluster:

Access graph

Apply intents

We will now declare that the client intends to call the server.

When the intents YAML is applied, creating a custom resource of type ClientIntents, Otterize will add a network policy to allow the intended calls (client server) and fail all unintended calls (e.g., client-other server).


You can click on the services or the lines connecting them to see which ClientIntents you need to apply to make the connection go green!

  1. Here is the intents.yaml declaration of the client, which we will apply below:
kind: ClientIntents
name: client
namespace: otterize-tutorial-npol
name: client
- name: server

See it in action

Keep an eye on the logs being tailed in the [client-other] terminal window, and apply this intents.yaml file in your main terminal window using:

kubectl apply -f

Client intents are the cornerstone of intent-based access control (IBAC).

as expected since it didn't declare its intents:

Calling server...
HTTP/1.1 200 OK
Hi, I am the server, you called, may I help you? # <- before applying the intents file
Calling server... # <- after applying the intents file
curl timed out
Calling server...
curl timed out
Not seeing the time out?

If client-other isn't timing out, then the installed CNI plugin likely does not support network policies. Consult the docs for your Kubernetes distribution or head back to the Calico installation section to install one. For example, Minikube does not start by default with a CNI that supports network policies but you can ask it to start with one that does, such as Calico.

  1. And in the [client] terminal you should see that calls go through, as expected since they were declared:
Calling server...
HTTP/1.1 200 OK
Hi, I am the server, you called, may I help you?
  1. You should also see that a new network policy was created:
kubectl get NetworkPolicies -n otterize-tutorial-npol

This should return:

NAME                                           POD-SELECTOR                                         AGE
access-to-server-from-otterize-tutorial-npol otterize/server=server-otterize-tutorial-np-7e16db 6s

If you've attached Otterize OSS to Otterize Cloud, go back to see the access graph in your browser:

Access graph

It's now clear what happened:

  1. The server is now protected, and is also blocking some of its clients. Click on it to see what to do about it.
  2. Calls from [client] are declared and therefore allowed (green line).
  3. Calls from [client-other] are not declared and therefore blocked (red line). Click on the line to see what to do about it.

Otterize did its job of both protecting the server and allowing intended access.

What did we accomplish?

  • Controlling access through network policies no longer means touching network policies at all.

  • Clients simply declare what they need to access with their intents files.

  • The next kubectl apply ensures that network policies automatically reflect the intended pod-to-pod access.

Expand to see what happened behind the scenes

Otterize generated a specific network policy on the ingress of the pods of a server, allowing the server to be accessed by the pods of a client. Otterize uses labels to define the network policy and associate it with a server in a namespace and a client in a namespace, as follows:

  1. The server's pods are given a label whose value uniquely represents that server. The network policy stipulates that it applies to the ingress of server pods with this label.
  2. The client's pods are given a label derived from the server's unique value. The network policy stipulates that only client pods with this matching label can access the server.
  3. The client's namespace is given a label whose value is the namespace of the client. The network policy stipulates that only client pods whose namespaces have this label can access the server. This is used to allow cross-namespace intents.

Otterize saved us from doing all this work by simply declaring the client's intents in intents.yaml, while the appropriate network policies were managed automatically behind the scenes.

Further information about network policies and Otterize can be found here.

Bonus tutorial

Try to create an intents file yourself for client-other, and apply it to allow this other client to call the server.

What's next


To remove the deployed examples run:

kubectl delete namespace otterize-tutorial-npol