Templates Reference

Templates define hardware characteristics and are reused across racks to avoid repetition.

Device Templates

Location: config/templates/devices/{type}/

templates:
  - id: bs-x440-a5
    name: "BullSequana X440 A5"
    type: server            # server, switch, storage, pdu
    role: compute           # optional: compute, visu, login, io, storage
    u_height: 10

    # Front-view layout (chassis node grid)
    layout:
      type: grid
      rows: 5
      cols: 4
      matrix:
        - [1, 2, 3, 4]
        - [5, 6, 7, 8]
        - [9, 10, 11, 12]
        - [13, 14, 15, 16]
        - [17, 18, 19, 20]

    # Rear-view layout (optional)
    rear_layout:
      type: grid
      rows: 2
      cols: 2
      matrix: [[1, 2], [3, 4]]

    # Health checks to run for this device type
    checks:
      - node_up
      - ipmi_temp_warn
      - ipmi_temp_crit

    # Metrics to display for this device type
    metrics:
      - node_temperature
      - node_power
      - node_cpu_load

    # Per-device tooltip thresholds (override the global metrics library defaults)
    display_thresholds:
      temperature:
        warn: 55    # °C — arc gauge turns amber above this
        crit: 70    # °C — arc gauge turns red above this
      power:
        warn: 400   # W
        crit: 600   # W

display_thresholds

The display_thresholds block customises the WARN / CRIT colour breakpoints shown in the HUD tooltip arc gauge, overriding the global defaults from the metrics library.

Key	Description
temperature.warn	Temperature (°C) above which the arc gauge turns amber
temperature.crit	Temperature (°C) above which the arc gauge turns red
power.warn	Power draw (W) above which the power bar turns amber
power.crit	Power draw (W) above which the power bar turns red

All keys are optional. Omitted keys fall back to the default thresholds defined in the metrics library YAML (config/metrics/library/).

When to use it: GPU servers and storage arrays typically run hotter than standard compute nodes. Set device-specific thresholds to avoid false-amber alerts on high-TDP hardware.

# GPU server — higher temperature thresholds than standard compute
display_thresholds:
  temperature:
    warn: 65
    crit: 80

Rack Templates

Location: config/templates/racks/

templates:
  - id: standard-42u-pdu
    name: "Standard 42U with PDUs"
    u_height: 42

    infrastructure:
      rack_components:
        - template_id: pdu-raritan-px3
          location: left     # left, right, front, rear
          id: pdu-left
        - template_id: pdu-raritan-px3
          location: right
          id: pdu-right

Rack Component Templates

Location: config/templates/rack_components/

templates:
  - id: pdu-raritan-px3
    name: "Raritan PX3 PDU"
    type: pdu              # pdu, switch, hmc, rmc, cooling
    location: side         # side, rear, front
    u_height: 1

    checks:
      - pdu_status

    metrics:
      - pdu_active_power
      - pdu_current
      - pdu_voltage

Template Types

Type	Description	Views
server	Single or multi-node server	Front/rear rack views
switch	Network switch	Front/rear rack views
storage	Storage array	Front/rear rack views
pdu	Power distribution unit	Rack component panel
hmc	Hardware Management Controller	Rack component panel
cooling	Cooling unit	Rack component panel