Data Center Commissioning Checklist for Owners: Prevent Rework Before IST Starts

Introduction
Integrated Systems Testing (IST) is the final exam for a new data center – when every backup generator, cooling unit, UPS, and control system must work together under simulated real-world conditions. It’s the culmination of months (or years) of design and construction, and a failed IST means the facility can’t go live. At that late stage, discovering a problem can set schedules back by weeks and cost a fortune in rework. In contrast, a smooth IST validates that all systems meet the design intent and owner’s requirements, allowing a timely handover to operations. This high stakes “all systems go” moment is why careful planning and coordination before IST is so critical. Even small gaps in earlier planning can turn into costly rework or delays during commissioning (www.opal-rt.com). The good news is that with the right approach, owners can catch and fix issues well before the integrated test – keeping the project on schedule and avoiding unpleasant surprises.

In this post, we’ll outline a data center commissioning checklist for owner teams to prevent rework before IST starts. These steps focus on aligning all stakeholders early, maintaining an up-to-date source of truth, and rigorously verifying systems at each phase. By following this checklist, hyperscalers and neocloud providers building data centers can ensure their facilities sail through IST on the first try, avoiding the scramble of last-minute fixes. We’ll also highlight how modern tools (like ArchiLabs’ cross-stack platform) help tie everything together – from design data to test scripts – to streamline the process. Let’s dive in.

1. Engage Commissioning Early and Define Clear Requirements

One of the biggest mistakes is treating commissioning as an afterthought. Instead, make commissioning a part of the project from day one, during design and planning. Start by defining the Owner’s Project Requirements (OPR) clearly – this document spells out exactly what the data center is supposed to achieve (capacity, redundancy, efficiency, uptime criteria, etc.). A strong commissioning process will tie every test back to the OPR to confirm the facility meets those targets (www.opal-rt.com). Equally important is completing a Basis of Design (BOD) that translates those owner requirements into specific design criteria. By engaging a commissioning team or agent early, you can review the designs against the OPR/BOD and catch misalignments or omissions long before equipment arrives on site. Early commissioning involvement means things like critical maintenance access, test points, and sensor placements are built into the design, not discovered as missing during IST.

Industry best practices (such as ASHRAE Guideline 0) break the commissioning process into multiple levels or phases from factory acceptance tests all the way to integrated system testing (www.dataknox.io). As the owner, insist on a documented Commissioning Plan that outlines these phases, the test procedures, and acceptance criteria for each level. Make sure this plan defines “gates” or entry/exit criteria for each stage – e.g. all Level 3 tests (equipment startup and pre-functional checks) must be completed and signed-off before Level 4 (system functional tests) begin. Adopting a gated approach with clearly defined milestones forces the project team to resolve issues early instead of pushing them down the line. In other words, you won’t start IST without proof that each subsystem met its prerequisites. This discipline dramatically reduces the chance of major faults emerging in the final hour. It’s wise to also plan integrating commissioning activities with the construction schedule – for example, scheduling load bank tests or redundancy failover simulations as part of the project timeline, not as an ad hoc exercise. By treating commissioning as a technical project within the project (with its own timeline, responsibilities, and quality benchmarks), you create accountability that reduces risk and rework (www.opal-rt.com).

Early planning should also cover risk identification and mitigation strategies. Ask your teams: what could delay or derail commissioning? Long-lead equipment, incomplete vendor documentation, late design changes, missing parts – all are common culprits. Mitigate these by tracking them in the project risk register and adjusting schedules or stocking spares as needed. For example, if replacement parts for a failing UPS take 12 weeks to arrive, you want to know that well before you’re supposed to start IST! By engaging in this level of planning and foresight, the owner’s team sets the foundation for a smooth commissioning process with minimal surprises.

2. Keep Design, Construction, and Commissioning in Sync (Documentation Control)

Data center projects evolve fast – designs get refined, equipment models get swapped, field conditions change. One major source of rework is mismatches between what’s built, what’s documented, and what’s tested. Owners can prevent this by enforcing strong documentation control and synchronization across all teams. In practice, this means every design change, submittal update, or field change order needs to be captured in a single source of truth that all stakeholders use. If the as-built reality drifts from the design on paper but test plans aren’t updated, IST may “fail” something that was actually intentional (or miss something that isn’t built as assumed). Prevent these scenarios by maintaining an always-up-to-date set of drawings, specs, and equipment lists throughout construction and pre-commissioning.

A practical tip is to hold regular coordination meetings specifically focused on upcoming testing. Bring together design engineers, contractors, and the commissioning agent to review any changes or issues before moving into the next phase of testing. For example, if a last-minute wiring change was made to a generator control circuit, ensure the commissioning scripts and diagrams are updated accordingly so the IST doesn’t flag it as a discrepancy. Many failures during integrated testing stem from lack of clear communication and documentation – something wasn’t installed per the latest design, or the design was revised but not everyone got the memo (www.prometheusgroup.com) (www.prometheusgroup.com). Avoid that by implementing robust change management: a formal process where any change in one system’s design or installation triggers evaluation of impacts on other systems and on the commissioning tests. Owners should require that updated documents (one-line diagrams, sequences of operations, layout drawings, etc.) are circulated and reviewed as living documents, not just frozen at 50% design.

Using a centralized data management tool can be a game-changer here. For instance, if you manage equipment inventories in Excel, cable layouts in CAD, and calibration data in a database, it’s easy for these to diverge. A centralized platform (more on this later) that connects your Excel sheets, DCIM, BIM models, and databases ensures that a change in one place propagates everywhere. The goal is that by the time you reach IST, every system configuration and setpoint in the facility matches what’s on the latest drawings and in the test documentation. With everything and everyone in sync, the integrated test won’t be derailed by an outdated spec or an incorrect wiring diagram. In short, rigorous document control and cross-team communication are your best defense against late-stage rework due to oversight.

3. Verify and Test Components Thoroughly (No Shortcuts on Pre-Functional Checks)

By the time you’re approaching IST, each individual component and subsystem in the data center should have been exercised and verified. The owner’s team should ensure that all earlier commissioning levels (factory tests, site installation checks, and system functional tests) have been completed successfully – and documented. It’s tempting under schedule pressure to rush into integrated testing, but skipping or shortcuting earlier tests is a recipe for failure. The IST is not the time to find out a backup cooling unit doesn’t start, or a generator’s auto-transfer sequence was wired incorrectly. Those should be caught in Level 3 or 4 testing, when you check each system in isolation. Insist on thorough pre-functional testing for every piece of equipment and every system (electrical, mechanical, controls, fire, security, etc.) according to the commissioning plan. This includes things like:

• Equipment startup and calibration: Verify every generator, UPS, CRAC unit, chiller, fire suppression system, etc. has been started up per manufacturer guidelines and any calibration/certification is done. For example, are temperature sensors and pressure gauges calibrated? Are protective relays on switchgear set to the specified trip settings? Calibrate and document these now so they don’t trigger false alarms later.
• Component redundancy and failover: Test redundant components in their failover modes. If you have N+1 UPS modules or A/B power feeds, intentionally take one offline to see that the load transfers seamlessly. If a backup CRAH unit should kick on when the primary fails, trigger that scenario in a controlled way. These functional tests prove each part does what it’s supposed to do on its own.
• Subsystem functional performance: Beyond individual devices, test subsystems as a whole. For example, run a power load test on each UPS+generator string to ensure it can carry the design load. Or do a cooling system test with all CRAHs and chillers under heat load to see that temperature setpoints hold. Check that control sequences (like generator start-> ATS transfer -> cooling interlocks) function correctly in each subsystem. Any anomaly here should be fixed now, not later.
• Integration of monitoring and controls: Ensure that your building management system (BMS) or DCIM is reading all the correct values and alarms. It’s much easier to fix a missing sensor reading or incorrect alarm mapping when you’re testing one system at a time than during the chaos of an all-systems test. Confirm that every device is reporting to the central dashboards and the alerts are configured as expected. This way, during IST, you can rely on your monitoring to give an accurate picture of system status.

For every test performed, demand a written record of the results and any issues found. A robust QA/QC process uses standardized checklists and test forms to ensure consistency (www.prometheusgroup.com). As the owner, you (or your commissioning authority) should review these test reports. Look for patterns or repeated problems – are multiple breakers tripping unexpectedly? Does a certain HVAC unit show performance below spec? Revisit these before proceeding. If all components and subsystems have passed their individual tests and been fine-tuned, the integrated test will mostly be about verifying combined operation, not debugging basic errors. Thorough upfront testing drastically limits the variables during IST, making it far more likely to succeed on the first run. It also builds confidence: the operations team will feel better seeing that each UPS and chiller has already proven itself in earlier tests.

4. Track Issues and Fix Them – Before IST

Commissioning is all about finding and fixing problems, but the key is fix everything early. Every test from factory through pre-functional will likely surface some deficiencies – maybe a wiring error, a failed part, a programming bug in the controls, or a construction mistake. These are normal; what’s not acceptable is carrying unresolved issues into the integrated systems test. Keep an issue log throughout the project and make sure there is a clear owner and resolution plan for each item. The owner’s team should regularly review this punch list with the contractors and engineers to push issues to closure. By the time you’re ready to start IST, that log should be burned down to only minor, non-critical items (preferably zero issues at all). If a serious problem is still open – say a generator fuel pump is acting up, or a cooling unit valve is stuck – do not start IST with that hanging. Pause, fix it, re-test the affected subsystem, and then proceed. It’s much more efficient to resolve it in isolation than to fail the whole IST and have to reschedule a repeat after repairs.

A helpful practice some mission-critical teams use is a visual tagging system for commissioning status. For example, each major piece of equipment might get a colored tag or label once it passes certain tests (www.dataknox.io). Red tags could mean pre-start checks incomplete, yellow for pre-functional tested, green for system functional tested, and so on. By the time you reach IST, every critical component should have the “green light” indicating it has passed its earlier verifications. If anything is still red or yellow, you know at a glance it’s not ready. This simple method ensures nothing slips through the cracks. It also fosters accountability – everyone can see which contractor or team is responsible for the remaining fixes (since tags often include the responsible party’s sign-off). Whether through tagging or just diligent documentation, strive for complete transparency on project readiness.

Communication is crucial here as well. Make sure that any issue discovered in testing is immediately communicated to all relevant teams. For instance, if an electrical test found a panel feed incorrectly landed, let the mechanical team know in case there’s an interlock or dependency. If a software bug in the DCIM caused an alarm failure, inform the controls team and the operations folks who will use that system. Often, what seems like a single-discipline issue can have ripple effects. An open issue log visible to all teams (potentially managed in a shared software tool) helps avoid silos. Ultimately, the owner’s role is to champion a culture of “no loose ends” going into IST. Emphasize that schedule pressure is not an excuse to leave things unaddressed – a rushed IST redo later will waste far more time than a small delay to fix something properly now. By holding this line, you’ll prevent the dreaded scenario of an IST abort followed by weeks of forensic troubleshooting. Instead, you’ll enter the integrated test with confidence that every known problem has been resolved, and any remaining surprises will be minimal.

5. Conduct a Pre-IST “Dry Run” and Prepare the Team

When all individual commissioning stages are complete and issues resolved, it’s wise to do a dry run for IST. Think of it as a dress rehearsal for the big test. This can be as simple as a tabletop run-through or as involved as powering up load banks for a trial run of certain scenarios. The idea is to double-check readiness: are all instruments in place and working? Are test scripts finalized and understood by everyone? Do all stakeholders know their roles during each test event? Taking time for a preparatory run can uncover last-minute oversights. For example, you might realize a critical communication channel was missing (“Who has authority to call the utility to simulate a power outage?”) or that the load bank cables won’t reach the intended connection points without re-routing. Far better to sort those out now than during the official IST window.

Logistics and resources must be lined up before IST starts. Ensure that all required test equipment is on site and calibrated – this includes things like load banks (with sufficient capacity to simulate full IT load), fuel for generators, multimeters and infrared cameras for thermal scanning, data loggers or commissioning software for capturing results, and so on. Double-check that your temporary load connections (if using load banks or dummy loads) are properly set up, with the right breaker sizes and cable lengths, and reviewed for safety. It’s useful to have backup equipment as well (spare load bank or extra sensors) in case of failures. As the owner, you want confirmation that the commissioning team has these resources ready.

Align all participants on the test plan ahead of time. Commissioning is a team sport – it involves designers, contractors from every trade, facility operations folks, and often third-party specialists. Before IST day, hold a kickoff meeting to walk through each test step, roles and responsibilities, communication protocols, and most importantly the emergency stop procedures. Testing a live facility involves risk (you will intentionally cut power, simulate faults, etc.), so everyone needs to know how to hit pause if something unsafe or unexpected happens. Make sure the data center operations team is ready to support these tests and that they’ve been briefed on what to expect (for example, “We will kill power to Rack Row A at 14:00 and it will be running on UPS/generator for 15 minutes”). If the data center is adjacent to or integrated with any live environments, coordinate carefully to isolate them – you don’t want an IST on a new system accidentally impacting a production environment. In some cases, owners choose to invite an independent commissioning witness or peer reviewer at this stage to audit the test readiness – a fresh set of eyes can sometimes spot gaps in the plan.

Finally, ensure all test scripts and procedures are finalized and approved by stakeholders (www.dataknox.io). Every scenario – utility outage, cooling failure, fire alarm, network cutover, etc. – should have a written step-by-step procedure that’s been reviewed by the design engineers and owner reps for safety and completeness. When everyone signs off on the scripts beforehand, it prevents confusion during the actual test. As part of this, confirm that acceptance criteria are clearly defined for each test. For instance, if you simulate a utility power loss, is it acceptable for one UPS to overload to 90%? How long can it take for the generator to come online? Define these criteria now so that during IST you can objectively say “pass” or “fail” without debate. With people, equipment, and plans all set, the IST can proceed like a well-rehearsed performance. The better this preparation, the more routine (and uneventful) integrated testing will feel – which is exactly what you want.

6. Leverage Automation and a Single Source of Truth (How ArchiLabs Can Help)

Modern data center teams are increasingly turning to software platforms to manage the complexity of design, construction, and commissioning. Relying on spreadsheets, email threads, and scattered documents is not only inefficient but also prone to errors that cause rework. This is where a cross-stack integration and automation platform like ArchiLabs can make a huge difference. ArchiLabs is building an AI-driven operating system for data center design and operations that connects your entire tech stack – Excel sheets, asset databases, DCIM systems, BIM/CAD tools (including Revit), analysis software, and custom apps – into one unified, always-in-sync source of truth. For an owner commissioning a data center, this means every piece of data (from design drawings to equipment lists to test results) stays linked and up-to-date across all systems. If a change is made in a CAD model or a new asset is added in a DCIM tool, that information propagates everywhere automatically, eliminating the version mismatches that often plague large projects.

Beyond just data synchronization, platforms like ArchiLabs bring powerful workflow automation to streamline commissioning and other operational tasks. Teams can automate repetitive planning work such as generating rack layouts, cable pathway designs, and equipment placement plans – ensuring these are done faster and with fewer errors. More relevant to commissioning, ArchiLabs can automate large portions of the testing workflow. For example, the system can generate standardized test procedures for each piece of equipment or sequence being tested, pulling in the relevant data (setpoints, expected outcomes) from your design specs. It can then assist in running and validating tests by reading live data from sensors or BMS systems via API integrations, comparing results against expected criteria, and flagging any deviations instantly. As tests are completed, ArchiLabs helps track results and aggregate reports – so you can see in one dashboard which tests passed, which had issues, and all the accompanying data and logs. This eliminates the tedious manual collation of commissioning data and ensures nothing is overlooked.

Crucially, ArchiLabs acts as a single repository for all project knowledge – syncing specs, drawings, change orders, and even operational documents in one place for easy viewing, editing, and version control. No more digging through email attachments to find the latest one-line diagram or trying to reconcile multiple Excel lists of assets. Everyone accesses the same information in real time. The platform’s AI capabilities also enable custom “agents” that teams can train to handle end-to-end workflows. For instance, you could have an agent that automatically reads equipment calibration certificates (PDFs or data files) and updates the central database, or one that interfaces with your CAD model to place and number assets according to your standards, then exports the updated bill of materials to an Excel sheet. These agents can span across tools – reading/writing to Revit or other CAD platforms, handling IFC models, pulling data from external databases or REST APIs, and even orchestrating multi-step processes across your entire toolchain. In the context of commissioning, you might deploy an agent to orchestrate integrated testing: it could trigger a simulation in one tool, collect sensor readings from another, alert team members via chat when a test step is complete, and push the final sign-off status to your project management system. This kind of cross-stack automation is transformative – it not only saves time, but also reduces human error by ensuring every system is talking to each other.

For owner-operators managing multiple data center projects or iterations, a platform like ArchiLabs becomes a force multiplier. It enforces consistency in how commissioning tests are prepared and documented, captures institutional knowledge (so each new project learns from the previous), and provides real-time visibility into project status. Instead of frantically updating spreadsheets and emailing out revision clouds on drawings, your team spends time on higher-value analysis and decision-making. ArchiLabs essentially serves as the central nervous system for your data center project data, ensuring that when IST day arrives, all your tools – from design models to test equipment – are singing from the same hymn sheet. The result is fewer last-minute hiccups, a faster commissioning process, and a more reliable data center from day one. By investing in this kind of integrated automation, leading operators (especially hyperscalers building at massive scale) can shave weeks off deployment schedules and avoid costly errors that slip through cracks in disconnected workflows. In short, leveraging such technology is becoming a best practice to de-risk commissioning and operations through data synergy and smart automation.

Conclusion

Commissioning a data center is a complex, high-pressure undertaking – but with the right checklist and tools, owner teams can master it. By engaging commissioning early, keeping everyone aligned to the design requirements, rigorously testing each component, and fixing issues before moving forward, you set your project up for success. When Integrated Systems Testing begins with all groundwork laid, it becomes a validation exercise rather than an emergency troubleshooting session. Projects that follow these practices see fewer surprises, minimal rework, and on-time go-lives, even as their facilities get ever more advanced. They also hand over to operations a data center that performs as expected from day one, with a complete trail of documentation and baseline metrics.

For today’s hyperscalers and innovative cloud providers, scaling data center infrastructure quickly is a competitive edge – but not at the expense of reliability. That’s why a preventative approach to commissioning (and the use of platforms like ArchiLabs to unite the process) is so valuable. It ensures that speed and quality go hand in hand. As you plan your next facility, remember that an ounce of prevention in commissioning is worth a ton of cure. Follow the checklist, invest in the right processes and automation, and you’ll save your team from late-night fire drills when IST rolls around. Instead, you’ll be watching a well-orchestrated test confidently demonstrate the resiliency and performance of your new data center – the ultimate payoff for thorough preparation. Here’s to zero rework and a seamless start-up on your next mission-critical project.