> ## Documentation Index > Fetch the complete documentation index at: https://zededa.mintlify.site/llms.txt > Use this file to discover all available pages before exploring further. > ## Agent Instructions > Use the documentation index to find relevant pages, then read the linked Markdown before answering and cite the source URL. > Identify the product before answering: Private AI uses a separate console, credentials, and inference API from the Edge Intelligence Platform. Do not apply instructions from one product to the other. > Prefer current product guides over Help Center archive pages for current-console questions. Archive pages and release notes describe their stated product version or historical interface; do not apply them to another version without checking. > Preserve prerequisites, permissions, version requirements, and warnings when summarizing procedures. Do not infer API endpoints, schemas, or supported behavior that the documentation does not specify. # Release 18.6 > Edge Infrastructure Services GUI 18.6.3 and API 18.6.7, published February 25, 2026. **Published:** 2/25/2026 GUI: 18.6.3 | API: 18.6.7 ## New GA Features ### Email or SMS Notification for Password Expiration You can now get email or SMS notifications about your Edge Infrastructure Services password expiration. From your **User Profile > User Details**, under **Notifications**, select **Email** or **Mobile**. If you don't choose one, you will continue to only see the notification on the GUI when you log in. The settings are also available in ZCLI. ## Enhanced GA Features ### ZEDEDA Edge Kubernetes Service Cluster Updates Updates for ZEDEDA Edge Kubernetes Service Cluster include the following: * Cluster Import. * Cluster Kubernetes Version Upgrade. * Download Kubeconfig. * API for bulk Cluster Kubernetes Version Upgrade. ### Single-Use Installer Updates The Single-Use Installer is now available for generic ARM, OnLogic FR-201, and NVIDIA Jetson-based devices. For the latter, the user can pick from JetPack 5 or JetPack 6 based images depending on whether the platform is Jetson Xavier or Jetson Orin based. ### Custom Branding Updates For OEM customers that operate on a standalone Edge Infrastructure Services instance, additional branding capabilities have been added for increased customization. ### Show or Hide Wireless Config Based on Adapter Type When onboarding or editing a node, the **Add Wireless Config** checkbox is no longer available if you select a network adapter other than WLAN. ### Edge App Adapters Tab From the existing **Edge App Instances > APP\_INST\_NAME > Adapters** tab, you can see the following changes for the MAC Address, IP Address, and Access VLAN ID fields. Previously the fields displayed **Not Available** when they were not configured or not reporting. Now the fields display **Not Configured** when they're not configured, and they display the reported values when available. ### Marketplace Updates * The **Upload an App** modal in the **Edge Kubernetes > Kubernetes Marketplace** of Edge Infrastructure Services is revised to better explain the supported formats. * In **Marketplace > Edge Apps** and **Marketplace > Models**, the cards now display in a different color when updates are available in the global marketplace. * In **Marketplace > Edge Apps**, there is now a **Project** filter available and an **Include apps available to all projects** checkbox. The checkbox enables you to filter on apps where the scope is set to all projects, rather than just on a specific project. ## New Early Access Features ### AI Assistant This Early Access feature is enabled through feature flags to selected customers. If you want to opt in as a beta tester and provide feedback on the feature, you can contact support. On your **Dashboard**, you'll see a new quick-action card called **Ask AI**. When you click it, the AI assistant opens in full-screen mode. It has ten popular topics that cycle across the screen. You can click a popular topic or you can use the AI Assistant Prompting Guide to ask your own questions. * The response you get might be a summary. You can click **Regenerate** to get more details, such as links to full articles in the Help Center. * You can minimize the AI assistant, so that it does not take up the whole screen. * You can **Copy** or **Export** as a PDF for temporary reports. * You can **Upvote**, **Downvote**, and provide detailed **Feedback** on the results. * From the existing **Edge App Instances > APP\_INST\_NAME > Logs** tab or **Events** tab, you can now also get an AI **Summary View**. ### ZEDEDA Edge Kubernetes and Node Cluster Updates These updates are all available together in EVE-k v16.7 or greater: * You can now use the EVE-OS (Edge Virtualization Engine) config property of **k3s.config.override** to set additional Kubernetes configuration parameters or override default options for nodes. * You can now use the EVE-OS config property of **k3s.version** to define the Kubernetes version of node clusters without updating EVE-k. * You can now enable strict node scheduling for node cluster app instances. The app will only run if deployment succeeds for the specified node ID. This is useful if all hardware in the cluster is not the same. Cluster app instances with AppPinning enabled will not failover to other nodes but do still retain the data redundancy features of volume instances on Node Clusters. * You can now use the API for cluster creation to pass the manifest. You can provide either the URL to the configuration manifest or you can provide the config as gzip(base64(manifest\_content)). The manifest content will be passed to the cluster nodes in the corresponding configurations so they can join the ZEDEDA Edge Kubernetes Service cluster. ## Fixes * The **Prepare Power Off for Nodes** pop-up message in Edge Infrastructure Services has been revised to reflect the current behavior for halted edge app instances. * The colors are changed for light and dark mode. * After you change your password, you are now directed to the **Dashboard** after a successful login. * The access and access labels are now aligned for **Reports > Utilization**. * During node onboarding, the **Unused** Network Interface setting is now properly retained. * **IPv6** is now removed from the UI for Local Network Instance, as it is only supported for Switch Network Instance. * The Entitlements API now correctly provides the permissions that enable the GUI to display the **Node Clusters** section appropriately for all authenticated users with access to the enterprise. * If Edge Sync is not configured at the device level, it now falls back to project-level Edge Sync configuration. * Under certain conditions, updating a network rule could erroneously trigger an app purge. This has been corrected. * Resolved an issue where attempting to update a node with VLAN adapters configured resulted in an error that the VLAN logical label cannot be the same as the interface name during VLAN sub-interface creation. * Resolved an issue where stale Edge View TCP sessions could prevent new VNC connections by implementing session activity validation. * MinIO support has been added enabling Edge Infrastructure Services support for both Amazon S3 or S3-compatible endpoints. Requires EVE-OS 16.0 LTS. ## Known Issues * In rare cases, if an Edge Kubernetes App Flows deployment through the cloud interface fails, some pods might be left over in the cluster. You would need to delete the pods manually by logging into the node using Edge View. * When you import an Edge Kubernetes Service cluster by running the kubectl registration command on the cluster node, the node is successfully provisioned, but you do not see the details in the GUI at **Edge Kubernetes > Clusters > CLUSTER\_NAME > Status** tab. However, you can see it in the cluster dashboard. * The Security Note records the IP address from which the change was made, but sometimes it shows the backend's localhost IP. * Sometimes the apps that do not have the update available might also appear in the list when you filter by Update available as Yes in the Models page in the UI. * The total model count with filter update available Yes/No in the Models screen might mismatch. * EVE image download (Installer/Live) fails on Single EVE OS Installer for ARM devices. This occurs only on the Canary Stage cluster. * If the AI assistant backend is not set up in a cluster, the option to enable the feature will still be visible in the entitlements settings of the enterprise. If it is enabled but the backend is not set up, the prompt will fail.