Skip to main content
Version: latest

Enable and Manage Proxy Configurations

You can add and manage proxy configurations for a Private Cloud Gateway (PCG) deployed into an existing Kubernetes cluster. By default, a PCG deployed to an existing Kubernetes cluster does not have a proxy configuration. If your infrastructure environment requires a proxy configuration, use the instructions in this guide to add and manage proxy configurations for a PCG deployed in a Kubernetes cluster and to ensure workload clusters deployed through Palette inherit the proxy configuration from the PCG cluster.

info

Workload clusters deployed through Palette will inherit proxy configuration from the PCG cluster. The PCG is not used as a network proxy for deployed workload clusters and does not provide internet connectivity for the workload clusters. Individual workload clusters must have their own proxy configurations to access the internet.

Use the following steps to add and manage proxy configurations for a PCG.

Prerequisites​

  • A PCG is deployed into an active and healthy Kubernetes cluster. Refer to Deploy a PCG to an Existing Kubernetes Cluster for additional guidance.

    warning

    If you deployed a PCG through the Palette CLI, refer to the respective platform installation guide for instructions on how to configure proxy settings during the installation process through the CLI.

  • The kubeconfig file for the Kubernetes cluster where the PCG is deployed. The Kubconfig file is used to authenticate with the Kubernetes cluster and deploy the Reach service.

  • Admin access to the Kubernetes cluster where the PCG is deployed. The Reach Helm Chart will create a namespace, service accounts, and roles in the cluster.

  • The extract utilities zip and tar are installed in the system you are using to deploy the Helm chart.

  • Palette tenant administrator access.

  • Proxy configuration details, such as the proxy URL, port, and authentication credentials.

  • The deployed PCG must have network connectivity to the proxy server.

  • The Kubernetes cluster where the PCG is deployed must have the proxy configuration. This includes any Certificate Authority (CA) certificates that are required to authenticate with the proxy server. This step varies depending on the platform where the Kubernetes cluster is deployed. Some platforms, such as managed Kubernetes services, may require additional steps. Below is a list of helpful links to set up proxy configurations for some common Kubernetes platforms:

    • Amazon EKS

    • Azure AKS

    • vSphere

      warning

      This feature has only been tested on AWS, Azure, and VMware vSphere. If you are using a different cloud provider, contact our support team for additional guidance.

  • Download the Reach Helm Chart provided by our support team. The Reach Helm Chart is used to deploy the Reach service into the Kubernetes cluster where the PCG is deployed. The Reach service is used to manage proxy configurations for the PCG. Contact our support team to obtain the Reach Helm Chart.

  • cert-manager is not already deployed in the Kubernetes cluster where the PCG is deployed. The reason for this is that you will deploy cert-manager as part of the enablement steps.

    info

    If cert-manager is already deployed, you can continue to use the service but make sure you edit the deployment to use the - --feature-gates=AdditionalCertificateOutputFormats=true flag. Otherwise, the Reach service cannot manage the proxy configurations.

Enable Proxy​

  1. Open a terminal session and navigate to the folder where you downloaded the Reach Helm Chart zip file.

  2. Unzip the zip file you received from the support team.

    unzip release-*.zip -d palette

  3. Navigate to the release folder.

    cd palette/charts/release-*/

  4. Deploy cert-manager to the Kubernetes cluster where the PCG is deployed. Use the following command to deploy the cert-manager Helm Chart.

    helm upgrade --values extras/cert-manager/values.yaml \
    cert-manager extras/cert-manager/cert-manager-*.tgz --install

  5. Next, navigate to the reach-system folder and extract the Reach Helm Chart.

    cd extras/reach-system/ && tar -xvzf reach-system-*.tgz

  6. Open the reach-system/values.yaml file in a text editor. Fill out the following YAML fields with the proxy configuration details:

    Field NameDescription
    reachSystem.enabledSet this field to true to enable the Reach service.
    reachSystem.proxySettings.http_proxyThe HTTP proxy URL, including the port number.
    reachSystem.proxySettings.https_proxyThe HTTPS proxy URL, including the port number.
    reachSystem.proxySettings.no_proxyA comma-separated list of URLs that should bypass the proxy.
    reachSystem.proxySettings.ca_crt_pathThe path to the CA certificate file used to authenticate the proxy server. Make sure the CA certificate is in the PEM format. If you do not have a CA certificate, leave this field empty.

    info

    A note on the no_proxy field: The no_proxy field is a comma-separated list of URLs that should bypass the proxy. Depending on your environment, you may need to add additional URLs to this list. The list provided in the code snippet below is a generic list that should work for most environments. The IP address 169.254.169.254, is required to be added to the no_proxy list. Otherwise, pods will not be able to start up successfully.

    aks.io,.amazonaws.com,.azure.com,.capz.io,.cluster.local,.gcr.io,.hubble-system,.kube-system,.kvdb,.microsoftonline.com,.privatelink.eastus.azmk8s.io,.spectrocloud.com,.spectrocloud.dev,.svc,.windows.net,10.0.0.0/8,10.10.128.10,10.10.192.1/18,10.96.0.0/12,127.0.0.1,169.254.169.254,.privatelink.eastus.azmk8s.io,.azure.com,.aks.io,.capz.io,.spectrocloud.com,gcr.io,windows.net,.windows.net,.microsoftonline.com,.amazonaws.com,.gcr.io,192.168.0.0/16,gcr.io,kubernetes,localhost,portworx-service,prometheus-operator-prometheus,windows.net

    The following is an example of a filled-out values.yaml file.

    reachSystem:
      enabled: true
      proxySettings:
        http_proxy: "http://172.16.0.4:1080"
        https_proxy: "https://172.16.0.4:1080"
        no_proxy: ".aks.io,.amazonaws.com,.azure.com,.capz.io,.cluster.local,.gcr.io,.hubble-system,.kube-system,.kvdb,.microsoftonline.com,.privatelink.eastus.azmk8s.io,.spectrocloud.com,.spectrocloud.dev,.svc,.windows.net,10.0.0.0/8,10.10.128.10,10.10.192.1/18,10.96.0.0/12,127.0.0.1,169.254.169.254,.privatelink.eastus.azmk8s.io,.azure.com,.aks.io,.capz.io,.spectrocloud.com,gcr.io,windows.net,.windows.net,.microsoftonline.com,.amazonaws.com,.gcr.io,192.168.0.0/16,gcr.io,kubernetes,localhost,portworx-service,prometheus-operator-prometheus,windows.net"
        ca_crt_path: "proxy-ca-cert.pem"

  7. Save the values.yaml file.

  8. Deploy the Reach service into the Kubernetes cluster where the PCG is deployed using the Reach Helm Chart. Use the following command to deploy the Reach service.

    helm upgrade --values reach-system/values.yaml reach-system reach-system-*.tgz --install
    WARNING: Kubernetes configuration file is group-readable. This is insecure. Location: /Users/ubuntu/.kube/config
    WARNING: Kubernetes configuration file is world-readable. This is insecure. Location: /Users/ubuntu/.kube/config
    Release "reach-system" does not exist. Installing it now.
    NAME: reach-system
    LAST DEPLOYED: Mon Jul 15 15:17:58 2024
    NAMESPACE: default
    STATUS: deployed
    REVISION: 1
    TEST SUITE: None

Once the Reach service is deployed, the PCG will use the proxy configuration in the Reach Helm Chart. If the cloud account is configured to use the PCG, the proxy configuration will be inherited by the workload clusters deployed through Palette.

Validate​

Use the following steps to validate that the Reach service is deployed and that the proxy configuration is enabled for the PCG.

  1. Open a terminal session.

  2. Issue the following command against the Kubernetes cluster where the PCG is deployed. Verify that the pods are active and healthy.

    kubectl get pods --namespace reach-system

    The output should display the Reach service pods in the Running status.

    NAME                                                READY   STATUS    RESTARTS   AGE
    reach-controller-manager-995c74db5-frb5j            2/2     Running   0          2m

  3. After configuring your cloud account to use the PCG, you can verify that the deployed workload clusters inherit the proxy configuration from the PCG. To verify the proxy configuration is inherited by the workload clusters, deploy a workload cluster through Palette. SSH into a node in the workload cluster and issue the following command to verify the proxy configuration is inherited.

    cat /etc/systemd/system/containerd.service.d/http-proxy.conf

    The output will display the proxy configuration details. Below is an example of the output.

    [Service]
    Environment="HTTP_PROXY=http://172.16.0.4:1080"
    Environment="HTTPS_PROXY=https://172.16.0.4:1080"
    Environment="NO_PROXY=.aks.io,.amazonaws.com,.azure.com,.capz.io,.cluster.local,.gcr.io,.hubble-system,.kube-system,.kvdb,.microsoftonline.com,.privatelink.eastus.azmk8s.io,.spectrocloud.com,.spectrocloud.dev,.svc,.windows.net,10.0.0.0/8,10.10.128.10,10.10.192.1/18,10.96.0.0/12,127.0.0.1,169.254.169.254,.privatelink.eastus.azmk8s.io,.azure.com,.aks.io,.capz.io,.spectrocloud.com,gcr.io,windows.net,.windows.net,.microsoftonline.com,.amazonaws.com,.gcr.io,192.168.0.0/16,gcr.io,kubernetes,localhost,portworx-service,prometheus-operator-prometheus,windows.net"