Six O&M Priorities That Define Reliable Data Center Infrastructure

Many facilities still rely on manual readings to judge whether equipment is “normal.” But the true insight lies in trend analysis. Equipment degradation usually develops gradually and often does not trigger alarms in the early stage.

Typical examples include:

• Temperature trends: Even within normal limits, a rise from 23°C to 26°C between two inspections may indicate localized cooling issues. A slow increase in return-air temperature over weeks often points to clogged filters or degrading heat exchange performance.
• Battery internal resistance: A steady upward trend, even below threshold, signals battery aging and warrants early intervention.

Trend-based inspection shifts the practice from data recording to predictive warning.

Sensors cannot capture everything. Human senses remain an irreplaceable “last line of defense.” Trained personnel can detect abnormalities through sight, sound, smell, and touch.

• Smell: 

Metallic odor may indicate poor electrical contact or arcing; burning smells often signal cable overheating; a sulfur odor may stem from battery overcharge; diesel smell could indicate fuel leakage.

• Sound:

Unusual vibration, humming, or friction noises from UPS units, CRACs, or pumps often relate to bearing wear, loose fasteners, or resonance issues.

• Visual cues:

Water marks, swollen batteries, abnormal condensation, or pooling water are early signs of underlying issues.

• Airflow sensation:

Humans can detect hotspot changes and airflow direction shifts that sensors may miss entirely.

Human-based inspection enables early-stage intervention long before faults surface in monitoring systems.

The period right after a service goes online or offline is one of the highest-risk windows. This is why international O&M standards commonly require joint inspection within 24 hours after any major change.

• After new workloads go live:

Validate power load distribution, cable temperature, PDU balance, local hotspots, and cooling adequacy.

• After workloads are decommissioned:

Verify resource release, equipment cleanup, and asset reconciliation to avoid “hidden assets” and residual risks.

Post-change inspection serves as a second validation layer, ensuring operational stability under the new configuration.

Single-device testing after upgrades or parameter adjustments may overlook integration risks. Inspections must confirm:

• Performance compliance: e.g., UPS efficiency, cooling performance of new CRAC units.
• Initial stability: New equipment often shows early-stage drift, vibration, or load fluctuations.
• Upstream/downstream impact: Compatibility, connectivity, and control logic interactions are common blind spots.

This step ensures upgrades do not introduce new vulnerabilities.

Inspections should focus on zones that carry high inherent risk or where sensors cannot fully cover conditions.

• High-risk areas: Fuel storage rooms, battery rooms, gas fire suppression zones, refrigerant storage, etc. These areas can lead to major incidents if overlooked.
• Monitoring blind spots: Cable tunnels, under-floor cavities, fiber pits, and similar locations. Even with sensors, coverage is often limited. Periodic field checks help identify water ingress, overheating, or blockages.

Such areas reflect the depth and professionalism of O&M practices.

Inspection priorities must adapt to seasonal environmental changes and be planned using historical data.

• Spring: Filter and fresh-air system checks to prevent clogging that impacts cooling.
• Summer: Focus on cooling performance, water circulation, and humidity control to prevent overheating and condensation.
• Autumn: Inspect drainage systems to avoid blockages during the rainy season.
• Winter: Check anti-freeze protection, static control, and heating system operation.

Seasonal inspection is a critical part of annual risk management.