Ashley Buckland, Managing Partner at JB Associates, explores the risks of relying too heavily on documentation and remote oversight when validating mission-critical infrastructure.
In the race to deliver new data centre capacity, commissioning is often viewed as just a final box to tick before handover. Documentation is completed, test sheets are signed off, and the project moves towards operation. But treating commissioning as an administrative exercise rather than a rigorous technical process presents a significant risk.
A data centre can be expertly designed, equipped with high-quality infrastructure and constructed by experienced contractors. Yet until every critical system has been tested, challenged and proven to operate together under real-world conditions, no one can be certain how the facility will perform when it is placed under pressure.
The view from the couch
That is why commissioning should never be managed solely from behind a desk. We sometimes see what we describe as ‘armchair commissioning’, where there is limited site presence and too much reliance on documentation provided by others. This creates risk because issues can be missed, tests may not reflect real operating conditions and problems may only become clear once the facility is live.
Effective commissioning demands a visible presence on site. It requires engineers who understand mission-critical infrastructure, question results, witness tests first-hand and verify that systems perform exactly as intended. Reviewing reports alone cannot reveal every issue. Being physically present during testing allows commissioning teams to identify anomalies, investigate unexpected behaviour and ensure nothing is overlooked.
The whole picture
This is particularly important in data centres, where resilience is the product being delivered. In a mission-critical facility, commissioning is not about demonstrating that individual pieces of equipment work. It is about proving that the facility behaves as designed when components fail, operating conditions change and systems are required to respond together.
Power distribution, UPS systems, generators, cooling infrastructure, building controls, fire protection and life safety systems must not simply work in isolation; they must operate seamlessly as a single integrated environment.
That is where integrated systems testing (IST) becomes indispensable.
Systems cannot be assumed to respond correctly to failures simply because individual components have passed factory or site acceptance tests. The loss of utility power, automatic generator start-up, UPS transition, cooling failures, control system responses and emergency operating sequences all need to be simulated, witnessed and validated under controlled conditions. Any weaknesses must be identified before customers begin installing live IT equipment, not after the facility has entered service.
Start earlier
Unfortunately, a heavy reliance on documentation, third-party reports and remote oversight can leave assumptions unchallenged and latent defects hidden until the building is carrying critical workloads. At that point, resolving issues becomes significantly more disruptive, costly and potentially damaging to the operator’s reputation.
In fact, successful commissioning begins long before equipment is energised. It starts during the design phase through commissionability reviews, the development of a clear commissioning strategy and early involvement from experienced specialists. It continues through factory acceptance testing, site testing, integrated systems testing, operator training and structured handover, ensuring every stage builds confidence in the next.
In simple terms, the most successful commissioning programmes share one common characteristic: they focus on proving operational performance rather than producing paperwork.
A recent project
We were recently asked to support a client on a data centre project that was approaching handover. The generator had been tested and the individual systems appeared ready, but when we ran the full IST on site, we identified an issue with the integration of the generator controls. It only became apparent when the systems were required to respond together as they would during a real event.
Had the client relied on individual test results and paperwork, the issue could have remained hidden until the facility was in operation. Instead, it was identified before handover, giving the team the opportunity to investigate, correct and retest it.
This illustrates why a test needs more than a pass-or-fail box. Before it begins, the team should agree on the operating conditions, the fault being introduced, the expected response and the evidence needed to demonstrate it. Everyone involved should know who will initiate the test, who is watching each system and when the test should be stopped if the response is unexpected.
This should all be made clear in pre-written test scripts, but the value of being on site is often in what happens between the planned steps. For example, a test sheet may record that a generator started, but those witnessing the sequence can see whether the controls gave the correct command, whether alarms appeared at the right time and whether the rest of the facility responded as intended. These details matter when a genuine fault occurs and the operations team must act quickly.
It is important that unexpected results are treated as useful findings, not inconveniences to work around to keep the programme moving towards handover. The cause needs to be fully understood and any corrective action retested as part of the complete sequence. Simply proving that one piece of equipment now works does not establish that it works with everything connected to it.
The end user or operator should also be involved. Watching these tests helps the people who will run the facility understand how it behaves during a fault, practise their response and challenge anything unclear in the procedures. Good commissioning leaves them with evidence that the facility performs as intended and the confidence to operate it when conditions are far from ideal.
Conclusion
Ultimately, a data centre is only as resilient as its commissioning process. A facility may appear complete, with every cable connected and every piece of equipment installed, but appearances can be deceptive. Until the building has been thoroughly tested, challenged and validated under realistic operating conditions, it cannot truly be considered ready for operation.
In mission-critical environments, confidence is earned through evidence, not documentation alone. Resilience must be witnessed, challenged and proven. And that cannot be done from an armchair.
