Cleaning is one of the few jobs that reaches every part of a live data hall. It includes the underfloor plenum, the ceiling above the racks, the cooling units, power distribution enclosures, cabling support infrastructure and the exterior of every cabinet. Each of these areas has equipment that can’t go offline, in a facility where planned work has to fit inside tightly controlled change windows.

Yet many operators still buy cleaning as a facilities service, scoped by square footage. As a result, it often sits outside the change management process that governs every other activity in the facility. That’s a gap. Cleaning a live hall is contamination control work. It carries real operational risk and deserves the same controls as any other work performed near energized equipment.

Where unmanaged cleaning creates risk

Anyone who has worked a raised floor can name the risks. Lifting floor tiles to reach the plenum releases the static pressure that pushes cold air up through perforated tiles. Open too many at once, or leave them open too long, and nearby racks lose cooling. Dust and particulates disturbed from plenums and overhead surfaces travel, and equipment intake fans pull them straight into servers.

Older raised floors add another hazard: zinc whiskers, which can become airborne during cleaning and cause intermittent equipment failures. Vacuums without HEPA filtration capture large debris while exhausting fine particles back into the room. Technicians working around cabinets can generate static discharge or brush against power cords, fiber runs and emergency power-off buttons.

Each of these risks starts with a person working without a procedure, or with one that does not fit the site. Uptime Institute's 2025 Annual Outage Analysis found that nearly 40 percent of organizations suffered a major outage caused by human error over the previous three years, and 85 percent of those incidents stemmed from staff failing to follow procedures or from flaws in the procedures themselves. The 2026 analysis found that failure to follow established procedures remains the leading driver of human error outages.

One unfortunate story that drives this home comes from a large financial institution who shared that their cleaning vendor systematically “turned the blinking lights” off when completing the task of detailing a rack’s contents. A technician unaware of their surroundings represents a significant risk when working in a live computing environment.

Treat cleaning like any other critical work

Data center teams already know how to manage risky work in a live environment. Every electrical or mechanical task on the data center floor runs through a method of procedure, a risk assessment and a change approval. Cleaning belongs in the same system. In practice, that means:

• A written method of procedure for each zone: Underfloor, overhead and plenum spaces, equipment, the environment and the top floor surface. Each one defines the sequence, tools, materials and limits technicians work within.
• Change approval: Every job goes through the site's change management process, given the same review that electrical and mechanical work receives. Operations staff assess the risk to live systems, sign off on the plan and schedule the work around change freezes, maintenance windows and peak load.
• Floor tile rules: A cap on how many tiles can be open at one time in a given area, with each tile replaced before work moves on. This cap should be governed by the static pressure in a given zone, typically between six to 10 tiles per zone.
• Trained technicians: Training for critical environments that covers static control, site badging, escort rules and the location of emergency power-off controls. A working understanding of airflow and power distribution along with cabling infrastructure is critically important when working in live environments. Properly trained technicians will also bring to light additional risks that are uncovered during a detailed cleaning process.
• The right equipment: HEPA-filtered vacuums, static-dissipative tools, low-lint wipes and cleaning agents approved for use around IT hardware.

Operators do not have to invent this structure. The cleanroom industry formalized it in ISO 14644-5, updated in 2025, which requires an operations control program covering written procedures, personnel training and a defined cleaning program for controlled environments.

Verify before the job closes

A cleaning job in a critical environment ends with measurement. Particle counts taken by zone will prove whether the space meets its cleanliness target and give operators a record to compare over time. ASHRAE TC 9.9 recommends that data centers be kept clean to ISO 14644-1 Class 8, which gives most operators a sound default target.

Those readings carry the most value against a baseline. A facility that verified its cleanliness at commissioning already has one, and a facility that did not can set one after an initial deep clean. Post-cleaning counts tracked alongside thermal and power data then show whether the environment is holding steady or drifting between service visits.

Manufacturers also write environmental requirements into their product documentation, and at least one major manufacturer states that environment quality "must be maintained over the lifetime of the system" for warranty and support, with remediation the customer's responsibility. Verified cleaning records are how an operator shows it met that obligation.

What every cleaning program needs

Whether cleaning is handled in-house or by a contractor, operators can hold the program to a short list of requirements:

• Written methods of procedure for each zone of the facility
• Change approval for every job
• A verified training program for all personnel entering the data center
• Defined equipment standards, including HEPA filtration and static control
• Floor tile limits written into the procedure
• Particle count verification after every job, recorded by zone.

Protect the baseline

Operators invest heavily in commissioning a clean facility. Construction debris is removed, the space is verified and the hall goes live in its cleanest and most stable condition. Ongoing cleaning determines whether it stays at its best. When run as critical maintenance, with procedures, approvals, trained technicians and verification, cleaning protects that baseline and removes a source of risk from the most sensitive room in the building. But when run as a janitorial task, it introduces one.

Learn more about Promera’s contamination control and critical-environment services for data centers, cleanrooms and life sciences facilities at www.promerasolutions.com.