Rack allocation workflow

  • Release version: Zurich
  • Updated July 20, 2026
  • 1 minute to read

    • The rack allocation agentic workflow reserves rack unit space in a datacenter by evaluating placement policies, capacity metrics, and change request requirements to find suitable rack allocations.

    Before you begin

    Role required: sn_ni_core.dc_ops_agent or sn_ni_core.inventory_agent

    The rack allocation agentic workflow reserves rack unit space in a datacenter by evaluating placement policies, capacity metrics, and change request requirements to find suitable rack allocations.

    To access the agentic workflow:

    1. Navigate to All > AI Agent Studio > Create and manage.
    2. Select Rack Allocation Workflow.

    Test the agentic workflow.

    To access the use case testing page:

    1. Navigate to All > AI Agent Studio > Testing.
    2. On the Overview page, select Rack allocation workflow.

    The Rack Allocation Workflow uses the Rack placement AI agent to execute allocation steps. The LLM matches the agent's name and description to workflow steps to select the appropriate agent.

    AI agents in the rack allocation workflow.

    Table 1. Rack allocation workflow AI agents
    AI agent AI agent role
    Rack placement AI agent

    Finds a datacenter rack allocation that fits a change request's capacity requirements.

    The agent runs autonomously based on the workflow instructions.

    Tools used by the agent

    • find_rack_placement — Finds racks that can accommodate the requested resources. Applies placement policies and capacity constraints. Returns a ranked list of suitable racks.
    • validate_change_request — Validates the change request. Verifies required fields are present and properly formatted. Returns validation status.
    • finalize_allocation — Finalizes the rack allocation flow: writes a work note summarizing the outcome and moves the change request to Design Complete state.
    • retrieve_change_request — Reads the change request from CMDB. Extracts target data center, RU, power, weight, temperature, and equipment requirements.
    • retrieve_policy — Retrieves placement policies from published Knowledge Base articles for the target datacenter and individual racks.