Using the knowledge hub

You can use the knowledge hub to upload and publish your organization's internal assets, such as guides, standard operating procedures (SOPs), and training materials. BMC AMI Platform makes published assets available for BMC AMI Assistant to use during chat interactions.

These assets enrich BMC AMI Assistant's knowledge base with your organization's internal expertise and context, which helps it provide more accurate, knowledge-based responses and supports more effective troubleshooting. 

Language support

The knowledge hub and BMC AMI Assistant support the following languages:

• English
• German
• French
• Italian
• Portuguese
• Spanish

Language support applies to both published assets and chat interactions.

Supported file specifications

Before you begin

To make sure of the proper knowledge hub integration, perform the following PID limit configuration after you deploy the LLM.

To configure PID limit (RKE2)

Set the node-level PID limit (run on all cluster nodes) as follows:

1. To backup, run the following command:
   sudo cp -a /etc/systemd/system.conf /etc/systemd/system.conf.bak.$(date +%F_%H%M%S) 2>/dev/null || true

2. To remove the systemd task limit, run the following command:

   sudo sed -i 's/^\s*#?\s*DefaultTasksMax=.*/DefaultTasksMax=infinity/' /etc/systemd/system.conf || true

   grep -qE '^\s*DefaultTasksMax=' /etc/systemd/system.conf || echo 'DefaultTasksMax=infinity' | sudo tee -a /etc/systemd/system.conf >/dev/null
3. To verify, run the following command:
   grep -nE '^\s*DefaultTasksMax' /etc/systemd/system.conf
4. To apply, run the following command:
   sudo systemctl daemon-reexec

Set the pod PID limit (RKE2) as follows:

1. To edit the RKE2 config, run the following command:
   sudo vi /etc/rancher/rke2/config.yaml
   1. You can either add or update with the following parameters:
      kubelet-arg:
      
        - "pod-max-pids=50000"
2. To restart RKE2, run the following command:
   1. For agent nodes
      sudo systemctl restart rke2-agent
   2. For server (master) nodes
      sudo systemctl restart rke2-server

To log in to BMC AMI Platform

1. In a web browser, navigate to the login URL in the following format: http://bmcamiplatform.example.com:30002.
2. Log in with a user who has knowledge hub permission. This is a user who is assigned to the ai_knowledge_admin role. For more information, see Roles and Permissions.

To add an asset

1. Click the Knowledge hub tab.
2. In the Assets window, in the right pane, click + New asset.
3. In the Add new asset window, on the File tab, click Attach files or drag your file to upload it. For supported file specifications, see Supported file specifications. To cancel the upload process, see To cancel an upload.
4. Specify who can use the asset in BMC AMI Assistant via the AI visibility controls:
   • Only me—BMC AMI Assistant can use this asset only within your chat interactions.
   • Everyone—BMC AMI Assistant can use this asset to provide knowledge in all users’ chat interactions.
5. Set email notification preferences. Email notifications let you continue your work while the platform processes the asset in the background.
6. Click Publish. 
   The publishing process uses machine learning services and might take some time, depending on system load. When you enable notifications, the system sends an email after the asset is published or if the publish operation fails. This removes the need to actively monitor the publishing process. The published files are displayed in the Assets table.
7. To view the latest status updates, click Refresh.

To cancel an upload

You can cancel the upload process at any time before publishing:

To manage an asset

In the Assets window, in the Assets table, click Action and select one of the following options to manage the asset:

To filter assets

To help you find assets efficiently, you can use the following filter types:

• Quick filters
• Advanced filter

Quick filters appear above the assets table as selectable categories, such as asset status or AI visibility. They provide a fast way to narrow down the list based on common criteria and to get an immediate overview of specific groups.

For more precise searches, the advanced filter helps you refine results by using detailed asset attributes such as type, status, creation and update dates, AI visibility, and asset owner. You can combine multiple conditions to focus on specific subsets of content, which is useful when managing large collections or reviewing assets at different stages of their lifecycle.

Asset statuses

The following table describes the statuses that assets move through during upload, publishing, and management: