Colocation Data Center Checklist: What to Verify Before You Move In

Signing a colocation agreement feels like the end of a long search. Your team has compared facilities, taken the tours, negotiated the monthly rate, and settled on a provider that seems to offer the power, space, and connectivity you need. But the decision that feels complete on paper is really just the beginning, because the facility you saw during the sales tour, with its spotless demo cages and rehearsed talking points, is rarely the facility you experience at 3 AM during an outage. The gap between the tour and the operational reality is where colocation relationships go wrong, and that gap only becomes visible after your hardware is on someone else’s floor and your options for correcting course have narrowed dramatically.

The stakes are higher than they appear, because a colocation move is not a transaction, it is a commitment measured in years. Contracts run 12 to 36 months, relocating racks of production equipment costs more than most teams budget for, and the monthly bill carries recurring charges for power, cross connects, and remote hands that quietly compound. Once your equipment is installed, every problem you failed to verify becomes an expensive problem to fix, because the facility now holds your hardware, your data, and your uptime as leverage in every negotiation. That is why a checklist matters so much: every item you verify before move in day is an issue you never have to resolve after it, when the cost of being wrong is highest.

Verify the Power Path End to End

Power is the most common reason teams regret a colocation choice, and the easiest to verify if you look beyond the headline number. The tour will tell you that racks are available with some kilowatt allocation, but the questions that matter are how that power is delivered, whether it is redundant, and how it is billed. Ask for the circuit configuration: how many circuits feed each rack, at what voltage and amperage, whether the facility can deliver 208 volt or 240 volt feeds for high density equipment, and whether the PDU is metered or monitored. Metered PDUs report current per phase, monitored PDUs report power per outlet, and the difference determines whether you can see actual draw or only total load. Ask to see the breaker schedule for the exact rack you will occupy, because a rack that has power available may share a circuit with a neighboring tenant whose equipment consumes the headroom you were promised.

Redundancy is where facilities differ most, so verify it at every level of the chain. Confirm that your rack has A and B feeds from separate distribution paths, that each feed is backed by its own UPS, and that the generator can carry the full facility load rather than only a portion of it. Ask about generator runtime and fuel delivery contracts, battery string replacement schedules, and the facility’s documented maintenance procedures for the power chain, because a redundant design that is never maintained is not redundancy, it is a failure waiting for a convenient moment. Ask whether the facility runs regular load tests on its generators and UPS systems, and whether it will share the most recent test reports, because a generator that has not been run under load since commissioning is a generator whose performance nobody actually knows.

Finally, ask how power is billed, because the billing model shapes both your monthly cost and your capacity planning. Per circuit billing is predictable but can leave you paying for capacity you do not use, while metered billing tracks actual consumption but requires accurate metering infrastructure in place. Whatever the model, ask for the rate card in writing before you sign, including the overage rate and the threshold at which it applies, because power overage charges are a frequent source of billing disputes after move in. Your own planning matters here too: document the expected draw of every device that will occupy the rack, using measured consumption rather than nameplate ratings, so the kilowatts you request from the facility match the kilowatts your equipment actually needs.

Cooling and Environmental Guarantees

Your equipment’s operating life and reliability depend on the thermal environment, and colocation facilities vary widely in how precisely they control it. Ask for the guaranteed temperature and humidity ranges in the SLA and compare them against your equipment’s operating specifications, because a facility that promises a comfortable range for typical servers may not support the density your high performance equipment needs. The industry reference is the ASHRAE thermal guidelines, which define recommended and allowable operating envelopes for different equipment classes, and a facility that commits to those envelopes in writing is making a stronger promise than one that offers a vague comfort range. Ask about hot aisle and cold aisle containment, the cooling capacity per rack and per zone, and what happens to your environment during cooling maintenance, because a facility that serves chilled water to your zone is a facility with a cooling architecture you need to understand before you depend on it.

Because cooling zones are shared, ask how the facility handles tenants whose heat output exceeds their allocation, since your inlet temperatures depend on what your neighbors do. Verify that the facility can show you current temperature and humidity readings for your cage area, that those readings are recorded and reviewable rather than visible only on a wall display, and that leak detection is present in the data hall, because water in a server room is one of the most destructive events a facility can experience. The density question deserves special attention: a facility whose standard allocation is 3 to 5 kilowatts per rack may be perfectly adequate for traditional servers and entirely wrong for GPU workloads, and the difference between the facility’s standard density and your actual density is a gap you should identify before move in rather than after. Ask what it takes to get additional cooling for a specific rack, whether that is a simple zone adjustment or a construction project, because the answer tells you how flexible the facility is about your growth.

The efficiency conversation is worth having even when the guarantees look adequate, because efficiency determines what your power bill looks like over the life of the contract. Ask about the facility’s power usage effectiveness and how it has trended, because a facility that publishes an improving number is investing in its plant while a facility that cannot answer the question may be running a cooling system that wastes energy in ways you will pay for indirectly. Ask whether cooling capacity is allocated per rack or per zone, and whether the facility’s monitoring systems can give you temperature and humidity data through a portal or an API rather than a phone call, because the more visibility you have into the environment your equipment lives in, the earlier you will notice the drift that precedes most thermal problems.

Connectivity and the Meet Me Room

The network connectivity you can buy at a colocation is the reason many teams move to one in the first place, so the interconnection ecosystem deserves as much scrutiny as the power. Start with carrier density: how many carriers and internet service providers are present in the building, whether the facility has a neutral meet me room, and how easy it is to add a new carrier after move in. A facility that locks you into a single provider or charges heavily for every new cross connect limits your networking options for the length of your contract, and those terms matter more than the headline bandwidth price. Ask for the cross connect pricing and typical installation turnaround in writing, because the difference between a facility that provisions a cross connect in a day and one that takes weeks is the difference between flexibility and frustration when you need a new link.

Diversity is the second connectivity question and the one that most affects real uptime. Ask how many separate building entrances the fiber plant uses, whether the facility’s network paths are truly diverse or share a common conduit, and whether your carrier connections can be physically routed through different entrances. A facility served by a single fiber entrance looks identical on paper to one served by multiple entrances, until a construction crew severs the only conduit and the redundancy you bought with dual circuits disappears because both leave the building through the same hole. Check latency to the locations your workloads depend on, whether that is your office, your cloud region, or your other data centers, because a facility that is close on a map can still sit on the wrong side of a congested peering point.

Beyond the traditional carriers, ask about the ecosystem that matters for modern workloads. Does the facility host cloud on ramps or direct connections to the major cloud providers, and at what cost? Can you get dark fiber or dedicated wavelengths into the building if your bandwidth requirements outgrow standard internet circuits? Is there a fabric or market exchange that lets you peer with other tenants, or is every connection a paid cross connect? These options determine whether the facility can grow with your network or whether you will hit a connectivity ceiling mid contract, when moving again is the only way out.

Ask about how bandwidth is sold in the facility, because the pricing models differ enough to change your monthly cost by a meaningful margin. Some facilities sell internet bandwidth as a committed rate with burst allowances, others meter usage, and others require you to buy transit through their carriers at their markup. Ask whether you can bring your own network equipment and terminate carrier handoffs directly into your cage, or whether the facility requires you to use its edge equipment and cross connects. Ask about port costs at the meet me room, whether the facility charges per cross connect or per port, and whether there are recurring fees for maintaining an unused cross connect that you will forget about until the bill arrives. Every recurring line item you identify during the sales process is a line item you can negotiate before you sign, and every one you miss is a line item you will discover on the first invoice.

Security and Access Control

Physical security is the easiest thing to under verify, because the sales tour is designed to make a facility look secure, and the difference between looking secure and being secure only appears in the details. Ask how access is controlled at each layer, the building perimeter, the data hall, and the individual cage, and whether each layer uses the same credential or requires separate authorization. Ask about the CCTV system, including camera coverage, retention period, and who can review the footage, because footage that is deleted after thirty days is useless for investigating an incident that surfaces months later. Ask about escort policies for visitors and contractors, whether tenant guests must be accompanied at all times, and how the facility handles tailgating. These details determine whether the security you are paying for actually protects your equipment or simply gives you a receipt for a promise.

The cage itself deserves inspection beyond the brochure photos. Confirm whether you are getting a lockable cage or a lockable rack, whether the cage mesh and doors meet your own security requirements, and whether you can install your own access control if the facility allows it. Ask about the facility’s process for revoking access when an employee leaves your company, because a facility that requires days of notice to remove a badge is a facility that keeps a credential active long after the person should have lost it. Ask whether you can audit the access log for your cage, and how long the facility retains those records. If your workloads carry compliance obligations, ask which certifications the facility holds and whether you can obtain a copy of the most recent audit report, because a facility that cannot produce its own compliance documentation will struggle to support yours.

Ask how the facility responds when security matters go wrong. What is the process if a badge is lost, if an unauthorized person is found in the data hall, or if a tenant reports a theft? How quickly are tenants notified of a security incident, and does the facility have an obligation to disclose events that affect their cages? Ask about the facility’s own employees: are they background checked, is access to the data hall restricted by role, and does the facility audit its own staff’s access records? A facility that treats security as a service it provides to tenants, with documented processes and audit rights, is a very different partner than one that treats security as a feature it sells in the brochure, and the difference shows up exactly when you need the security to be real.

Read the Contract Terms Like an Engineer

The SLA in your colocation contract defines what the facility owes you when things go wrong, and most teams discover too late that their SLA protects the facility more than it protects them. Read the uptime commitments for power, cooling, and network connectivity separately, because a facility can offer a strong aggregate uptime number while the specific components your equipment depends on carry weaker guarantees. Ask how service credits are calculated, whether credits are paid automatically or require you to file a claim, and whether caps or exclusion periods make the credits meaningless in practice. A credit is only compensation if it is automatic and material, because an SLA that requires a written claim within five days and caps credits at ten percent of one month’s bill will cost you far more in downtime than it ever returns.

The non SLA terms are where the real costs hide. Ask about maintenance windows: how much notice the facility gives for planned maintenance, whether maintenance happens during defined windows or whenever the facility chooses, and whether your A and B power paths are maintained independently so that maintenance never takes down both. Ask about remote hands availability, the hourly rate, the response time, and whether staff are on site around the clock or only during business hours. Ask about power overage rates, cross connect fees, and every line item that can increase the monthly bill, and ask the facility to confirm in writing that the quoted rate is the rate you will be charged for the full contract term. Ask about the escalation path when something fails, who you call, how fast the response obligation is, and what happens if the facility misses it. Finally, ask about the exit: termination notice periods, decommissioning fees, and what happens to your equipment if you stop paying, because the cost of leaving a colocation is part of the total cost of choosing it.

Two clauses deserve a careful read even though they rarely appear in the sales materials. The first is force majeure, which defines what happens when events outside the facility’s control interrupt your power or connectivity, because the scope of that clause determines whether you can claim anything at all when a regional event takes the building down. The second is insurance, both the facility’s liability coverage and what the contract requires you to carry, because some facilities require tenants to name them as additional insured or to carry coverage levels that are easy to miss during negotiation. Ask whether the facility’s liability insurance covers damage to tenant equipment caused by facility infrastructure, such as a cooling failure or a power event, because the answer determines whether your recovery comes from their insurer or yours.

Inspect the Floor Yourself

A scheduled tour shows you the facility at its best, so plan a second visit that treats the tour as a starting point rather than the evidence. Walk the exact space you will occupy and verify the cage or rack dimensions with a tape measure, because the rack in the sales material may be a half rack or a rack with a column obstruction that reduces usable space. Check the floor, confirming that the raised floor rating or slab can support your planned equipment weight, and look for signs of real operating conditions such as high density zones running hot, unlabeled cables, or cages that sit empty because tenants left. Test what you can: verify that the power outlets in your space are live and correctly labeled, confirm that the cross connect patch panels are where the facility says they are, and check the cable pathway from your cage to the meet me room, because a cage with no accessible cable path is a cage you cannot connect. If a tenant with similar equipment to yours exists in the building, ask them about power stability, cooling, and staff responsiveness, because no tour will tell you what the people already paying the bills have learned.

The visit is also the time to check the operational details that only exist in person. Look at the loading dock and freight access, because moving your equipment in is a physical operation that depends on dock scheduling, lift availability, and the width of the hallways between the dock and your cage. Ask how equipment deliveries are handled, whether the facility has a staging area for inbound hardware, and whether there are restrictions on delivery hours. Check the state of the facility’s labeling and cable management in occupied cages, because a facility where existing tenants keep clean, labeled environments is a facility that enforces standards, while a facility where cages are tangled and unlabeled will not help you keep yours orderly. Take notes and photographs during the visit, and record the measurements and observations in your infrastructure platform so the space you verified becomes the basis for your rack plan rather than a memory that fades by the time the hardware arrives. Obelinf’s rack management feature is the natural place to capture what you find, turning the verified space into a plan for where each device will land.

Document Everything Before Equipment Arrives

The most expensive documentation mistake in a colocation move is waiting until move in day to start recording what you have. By the time your hardware is on the floor, the team is focused on racking, cabling, and powering equipment under time pressure, and documentation is the first task to be deferred, which is exactly when errors are most likely to happen. Start before the move: create the asset list for every device that will occupy the new space, plan the IP address management scheme for the facility, assign the VLANs, and record the circuit IDs and cross connects you have ordered so that the information exists before it is needed. A checklist is only useful when it is checked against reality, and the reality you check against has to be written down before move in day.

The records you create before the move become the baseline for everything that follows, so they need to be structured rather than scattered. Every device needs a home in a device inventory that records its rack position, its power feeds, and its network connections. Every cable needs a documented path from source to destination, which is where cable tracking matters, and every circuit needs a record that includes the provider, the bandwidth, the handoffs, and the contract terms. When the facility hands you a cross connect ticket or the carrier confirms a new circuit, that information should be recorded immediately, because the gap between a documented change and an undocumented change is where colocation records quietly go stale. Plan the labeling convention for the new facility before the move as well, because labels applied during installation, when the racks are empty and the cable paths are clear, are labels that actually get applied, while labels deferred to a later maintenance window are labels that never happen.

Plan the Move In Sequence

The move itself deserves the same rigor as the contract, because the physical installation is where the plans you made either hold together or fall apart. Sequence the work: schedule the delivery with the facility’s dock, confirm the lift and hallway clearance from the dock to your cage, and stage equipment in an order that lets you rack the heaviest and deepest devices first, so you are not trying to lift a storage array over a row of already installed servers. Install the cable management and vertical PDUs before the devices, label every cable as it is pulled rather than after, and terminate the cross connects to the meet me room with enough slack that future changes do not require recabling. If you are using the facility’s remote hands for the installation, give them a written plan that includes the rack elevation, the device placement, and the power feeds each device should land on, because an installation crew works best when the plan is explicit. Obelinf’s free move in checklists and templates give that plan a head start, covering new site deployments, rack installations, and cross connect inventories in a format your team can use as-is.

Verification does not end when the last screw is tightened. Power on each device in sequence and confirm its management interface comes up, then test the redundancy you paid for: pull one feed and confirm the equipment continues to run on the other, then restore it and test the second feed. Confirm that cooling is keeping inlet temperatures where you planned, verify that the cross connects pass traffic at the negotiated bandwidth, and check that every device, cable, and circuit you documented before the move matches what is physically installed. This is the moment the checklist pays for itself, because any discrepancy found during move in day verification is still cheap to fix, while the same discrepancy discovered a month later, after the facility has added other tenants and your equipment has settled into its environment, is a problem that compounds. Record the results as you go, so the records you built before the move become the verified baseline for everything after.

Your Colocation Records, Ready Before Day One

Everything in this checklist, from the power chain to the cross connects, produces information your team needs to act on, and the teams that manage colocation well treat that information as a system rather than a pile of documents. Obelinf gives you a single platform where the facility, the cage, the racks, the devices, the cables, the circuits, and the IP space all live as connected records, so the checklist you work through before move in day produces data in the same place you will operate from after it. Each rack carries its position, its devices, and its power context, each device links to its interfaces and cables, and each circuit links to its handoffs, which means the questions this checklist asks become answers you can see at a glance instead of facts you have to hunt for. The same platform that holds your site records also holds the carrier contracts, cross connect IDs, and power details you negotiated, so the terms you verified during the sales process stay attached to the infrastructure they describe.

The value of that structure compounds after move in. When the facility calls about a maintenance window, you can see which devices sit on the affected feed and what they connect to. When a carrier asks about a circuit, the contract and handoff information is one click away. When you plan the next expansion, the utilization of your power, space, and connectivity is already visible, because every move in day detail was recorded as it happened. Obelinf keeps your data center management records accurate from the day you sign the contract, so the checklist you complete before you move in is the same source of truth your team relies on for every decision after it. Sign up at obelinf.com and start your colocation planning with a system that keeps every verification, every circuit, and every device in one place.