Cooling decisions in a data hall are only as good as the temperature data behind them. Many sites monitor return air at the cooling units and assume the IT equipment is comfortable, which hides hot spots and leads to over-cooling everywhere else. This checklist sets out what a sound thermal monitoring scheme should cover. | 1. Measure what the IT equipment breathes. Measure air temperature at the inlet of the IT equipment, because that is the condition the hardware manufacturers specify. | Use the recommended envelope in the ASHRAE TC 9.9 thermal guidelines, 18 to 27 °C at the inlet for most equipment classes, as the operating reference, and record humidity against the same guidelines. | Place sensors at the top, middle and bottom of the rack inlet. Temperature often rises towards the top of the rack, especially where hot air recirculates over the cabinets. | 2. Cover the whole airflow path. Supply air temperature at each cooling unit or in the floor void. | Return air temperature, to calculate the temperature difference across each unit. | Hot aisle temperature where containment is installed, to confirm that it is working. | Differential pressure across the containment or the raised floor, which shows bypass and recirculation before temperatures move. | 3. Make the data usable. Log at intervals short enough to capture transients, for example every minute, and keep at least a year of history. | Bring the readings into the BMS or DCIM platform alongside cooling plant status and IT load, so that cause and effect can be seen together. | Set alarms on the rate of change as well as on absolute limits; a fast rise often matters more than a value. | 4. Act on the readings. Close gaps with blanking panels and brush strips where inlet readings show recirculation. | Raise supply air temperature in small steps while inlet readings stay inside the recommended envelope, and record the energy effect. | Review the sensor layout whenever racks are added, moved or re-densified. | 5. Align with the standards. BS EN 50600-2-3 covers environmental control in data centre facilities and is a useful reference when writing the monitoring requirements into a specification. For energy reporting, a well instrumented hall also makes it much easier to calculate PUE correctly over a full year. | Next steps. Read our worked method on using PUE to size a cooling upgrade, or contact our team for an independent thermal review of your data hall.