Cloud customer?
Start for Free >
Upgrade in MyJFrog >
What's New in Cloud >


The Google GCP GKE Kubernetes App allows Google's GCP users to deploy a high availability instance of JFrog Artifactory (and optionally Xray) with a few keystrokes. This deployment is supported for GCP's GKE and via Anthos (for Hybrid/On-Prem) GKE. 


  • Active GCP subscription

  • Artifactory Enterprise Licenses (2+) 

    For Xray, enterprise licenses with Xray add-on and trial licenses are required.

  • A domain name with corresponding SSL Certificate (strongly recommended for production).

Page Contents


Setting up Artifactory Using the Google GCP GKE Kubernetes App

  1. Go to JFrog Artifactory Enterprise listing.
  2. Select Configure and select the target project to deploy.
  3. Select the cluster you wish to deploy to or create a new one.
  4. Enter all required details, and create a new master key. 
  5. Select Deploy. 
    The application is deployed.
  6. Follow the instructions on the right hand side of the App to obtain the public IP.
  7. Upload the license via the UI or REST API.

Upgrading Artifactory in the Google GCP GKE Kubernetes App

The upgrade is only supported for 6.x to 6.x and 7.x to 7.x with an expected small downtime window.

Go to the GKE workloads page, select the project where the app is deployed.The workloads must be upgraded manually, according to the following steps.

  1.  Select the master workload: <app-name>-artifactory-ha-primary master.
  2. Change YAML from<existing version>  to<new version>  in all locations, without changing the annotations.
  3. Select the JFrog GKE registry from a list of all versions (optional).
  4. Change the image to pull directly from JFrog, where images are updated at a more regular interval. Use<new version> (optional). 
    1. Scale the deployment to zero.
    2. Increase the deployment to 1.  
    3. Save the change.
      The new version of the primary workload appears.
  5. Change the image to pull from the secondary workload: <app-name>artifactory-ha-member.
    1. Scale the deployment to zero.
    2. Increase the deployment to previous member count (usually 2).
      The new version of the secondary workload appears.
  6. Verify all the nodes possess the correct version of the UI | Administration | Monitoring | Service Status.
  • No labels
Copyright © 2022 JFrog Ltd.