Do not approve a colocation lease because a facility advertises a Tier level, N+1 equipment, or an uptime percentage. Approve it only after the operator demonstrates that the exact space, power path, cooling envelope, maintenance procedures, and contract remedies support your equipment and business risk.
Start with your required IT load and environmental limits, then trace the facility’s utility-to-rack power chain, test its cooling assumptions against your planned heat load, walk through failure and maintenance scenarios, and turn every material promise into lease language. The evidence must apply to the proposed room, cage, suite, or rack—not merely to the building or the provider’s wider portfolio.
As an Amazon Associate I earn from qualifying purchases.
Set your acceptance criteria before touring a facility
A data center normally bundles space, electrical power, cooling, connectivity options, and physical security. Those services are not delivered in the same way everywhere. Before asking an operator to describe its design, document what your deployment actually needs. Otherwise, a sales presentation can make an adequate building appear suitable for an unsuitable load.
Recommended Free Tools
Document the IT load and growth plan
List the equipment that will be installed at move-in and the equipment expected during the lease term. Record nameplate input power, measured consumption where available, voltage and phase requirements, plug types, dual-cord or single-cord operation, inrush characteristics, and the highest credible simultaneous load. Distinguish between a cabinet’s nominal allocation and the load you expect to draw.
State the initial and target load in the units used by the proposal (for example, kW per rack or total kW). Ask whether the operator commits capacity as a fixed reservation, a metered allowance, or a shared pool. Define what happens if your measured demand exceeds the allocation: throttling, a surcharge, a required upgrade, a prohibition on adding equipment, or another remedy. These are commercial terms, not universal features of colocation.
Record environmental and availability requirements
For every equipment class, obtain the manufacturer’s allowable inlet-temperature and humidity range. Note any restrictions for storage, battery systems, accelerators, storage arrays, or other high-density hardware. Decide whether your business needs continuous operation during maintenance, protection from a single component failure, or simply a controlled shutdown after a prolonged utility event. That risk decision determines how much redundancy and contractual protection is justified.
Define evidence you will accept
Ask for current, site-specific documentation rather than a generic brochure. Useful evidence includes a one-line electrical diagram, cooling and controls diagrams, capacity calculations for the proposed footprint, maintenance procedures, test records, certification documents, and the technical schedules that will be attached to the lease. Mark which documents are informational and which obligations the operator will warrant contractually.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Trace the complete power path
The practical question is not “Does this building have generators?” It is “What sequence of equipment carries power from the utility to each of my power supplies, and what remains available when any part of that sequence is being maintained or has failed?” Ask the operator to walk the route with you using a current one-line diagram or equivalent record.
Identify every stage from utility to rack
The diagram and site walk-through should identify:
- Utility service entrances and the number of utility feeds.
- Service transformers, switchgear, breakers, and automatic transfer equipment.
- Uninterruptible power supply (UPS) modules, battery strings, bypass arrangements, and downstream distribution.
- Generators, paralleling gear, fuel storage, controls, and the connection point to the critical bus.
- Remote power panels, busways, floor PDUs, cabinet PDUs, receptacles, and any final-circuit monitoring.
- Components or paths shared by multiple tenants, rooms, phases, or buildings.
For each element, ask whether it serves your proposed space directly or through a shared distribution system. A second generator does not create an independent path if both generators depend on one shared switchboard or control system.
Understand what the UPS does—and does not do
ASHRAE describes UPS systems as isolating equipment from power disturbances and bridging the interval until generators start or an orderly shutdown can be performed. That makes the UPS a ride-through and power-conditioning system, not an unlimited energy source. Ask for the battery autonomy stated for the operating conditions relevant to your load, the low-battery and bypass behavior, and the process for an orderly shutdown if generator support is unavailable.
Have the operator explain the generator start and transfer sequence in plain language. Ask which signals initiate a start, how the critical load is transferred, what happens if a generator fails to start, and whether another source carries the load. Request the tested sequence and the date of the most recent test rather than accepting an unqualified “automatic” description.
Clarify fuel, testing, and maintenance responsibilities
Ask who owns the fuel, who monitors inventory, how replenishment is arranged during a regional disruption, and whether access restrictions or local rules affect delivery. Request the generator and UPS preventive-maintenance schedule, the types of load-bank or transfer tests performed, and the notification process for tests that affect your equipment.
Generator presence alone does not establish uninterrupted service. Utility failures, failed starts, transfer equipment faults, depleted fuel, controls problems, and a shared distribution failure can produce different outcomes. Your SLA should state which of these events are covered, excluded, or treated as planned maintenance.
Check metering and capacity enforcement
Determine where consumption is measured, how often readings are taken, and whether billing uses kW, kWh, amperage, a demand maximum, or a reserved-capacity fee. Ask whether the meter is at the rack, PDU, room, or upstream distribution point. Confirm the accuracy and dispute process for readings, the definition of a peak, and the notice period for an alleged overage. If the facility reserves capacity that you do not immediately use, identify when that reservation can be reduced or expanded.
Prove that cooling matches your equipment
Cooling suitability is a comparison between conditions at your equipment’s inlet and the equipment manufacturer’s allowable operating envelope. A room-level temperature claim is not enough: airflow patterns, rack density, humidity control, and the position of sensors can determine the conditions your servers actually experience.
Obtain the operating envelope
Request the permitted temperature and humidity ranges at the equipment inlet, the sensor locations used for alarms, and the assumptions behind those limits. Ask for the airflow arrangement (for example, the direction of supply and return air), the rack orientation expected by the design, and any blanking-panel or containment requirements. Compare those conditions with the most restrictive device in your planned deployment.
Match reserved capacity to heat load and density
Convert the committed electrical load into the heat load the cooling system must remove. Ask how much cooling capacity is reserved for your specific room, cage, or rack row and what rack density is permitted. Clarify whether the stated capacity assumes a particular outdoor temperature, economizer mode, chilled-water temperature, or equipment configuration that is not yet installed.
For high-density or liquid-cooled equipment, obtain the supported interface in writing. Confirm supply and return temperatures, flow or pressure limits, connection standards, leak detection, controls, isolation valves, and the capacity reserved for your footprint. Ask what happens to liquid-cooled equipment during maintenance or a cooling-system fault, including whether a temporary air-cooled operating mode exists and what limits apply.
Examine redundancy and failure behavior
Ask whether a single cooling unit, pump, chiller, cooling tower, control panel, or distribution path can be removed for planned maintenance while your committed load remains within the environmental envelope. Require the operator to describe the resulting airflow and temperatures, alarm thresholds, response ownership, and time allowed for load reduction.
Free tools Windows power users keep installed
One-click scans. No signup required.
ASHRAE’s Thermal Guidelines for Data Processing Environments, referenced in its handbook as the 2021 fifth edition, is useful background for discussing thermal conditions. It does not replace a comparison between your equipment specifications and the site’s contractual limits. ASHRAE also notes that temperature can rise temporarily around utility interruptions and restoration of generator-backed cooling; ask how long such a condition may last and what alarms or tenant actions are expected.
Ask about controls and monitoring
Identify which building-management or environmental data you will receive, at what interval, and whether historical records are available. Ask who can change set points, how alarms are escalated, and whether tenants receive notice when a cooling unit is placed in a degraded or maintenance mode. A dashboard without an agreed response procedure is not a reliability commitment.
Interpret N, N+1, 2N, and concurrent maintainability correctly
Redundancy labels are shorthand, not a complete description of your service. Require the operator to define each term for the exact space and topology proposed to you.
What N means
N is the number of components or capacity units required to carry the intended load under the stated design assumptions. If a system needs four pumps to support the load, four pumps represent N. The label says nothing about what happens during maintenance, a failed control, or a shared upstream fault.
What N+1 means
N+1 adds one additional component or capacity unit beyond the minimum required. The extra unit may allow one chiller, UPS module, pump, or generator to be unavailable while the remaining units carry the load. However, the arrangement may still use one electrical or mechanical path.
ASHRAE’s 2023 data-center guidance states: “The term N+1 is often used in the industry to describe a level of redundancy. However, this term alone does not provide sufficient detail to fully understand the level of redundancy provided, or to ensure that the resulting facility is actually concurrently maintainable.” Ask whether the spare component is connected to an independent path, whether controls and switchgear are also redundant, and whether the design remains valid in a maintenance state.
What 2N means
2N generally describes two complete capacity systems, each capable of carrying the required load. The useful question is whether the two systems remain genuinely independent all the way to your equipment. Shared utility service, switchgear, controls, fuel systems, distribution panels, or cable routes can create a common point of failure even where two major modules are shown on a diagram.
What concurrent maintainability means
A concurrently maintainable design is intended to permit planned maintenance on a component or path without shutting down the critical load. Ask the operator to identify the maintenance boundary, the switching procedure, the people authorized to perform it, and the conditions under which the claim no longer applies. Request a scenario-based explanation rather than relying on the phrase in a marketing document.
Run failure and maintenance scenarios with the operator
Use the same scenarios for every facility you are considering. Record the surviving equipment, expected alarms, operator response, tenant notification, and contractual treatment.
- Planned UPS maintenance: Ask which bypasses or alternate modules carry your load, whether power quality changes, and what work must be completed before redundancy is restored.
- Loss of one utility feed: Identify the transfer path, the loads that move, and any shared equipment that remains exposed.
- Generator start failure: Ask what starts next, how long the UPS can support the load, and who decides whether to shed or shut down equipment.
- Chiller or pump outage: Determine the remaining cooling capacity, the environmental alarm thresholds, and the maximum time before your load must be reduced.
- Transfer-equipment failure: Trace the effect of a failed automatic transfer switch or manual bypass and identify the work-around.
- Shared distribution failure: Ask what happens if a switchboard, busway, control panel, or common cable route serving multiple tenants is unavailable.
- Loss of cooling during electrical restoration: Confirm whether cooling is generator-backed, the startup order, expected temperature rise, and the tenant notification process.
For each scenario, ask whether the proposed redundancy survives the planned maintenance state. A system that tolerates a component failure in normal operation may not tolerate that same failure while another component is already isolated for work.
Use Tier classifications as evidence, not a guarantee
Uptime Institute’s Tier framework discusses facility capabilities related to maintenance, power, cooling, and fault conditions. A Tier document can help structure questions, but it does not replace site-specific due diligence or a contract review.
Verify the scope and status
Ask for the exact certification or validation document and confirm whether it covers the building, the phase, the room, and the services proposed to you. Check its current status and any limitations. A certification for a facility phase or design may not establish that later modifications, tenant distribution, or operating practices provide the same outcome.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Understand why component lists are insufficient
Uptime Institute’s Tier I description includes a UPS, dedicated cooling, and a generator while noting that preventive maintenance can require a full shutdown. That example illustrates why the presence of individual components does not prove continuous service. Your review must include maintenance procedures, staffing, switching controls, and the SLA’s treatment of planned work.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Translate technical claims into lease and SLA terms
Technical assurances have value only when the contract defines the service, measurement, exceptions, and remedy. ASHRAE recommends evaluating facility redundancy before signing and using a carefully worded SLA.
Define the covered service
State whether the commitment covers electrical power, cooling, environmental conditions, or only a broader facility availability service. Identify the precise delivery point and whether your equipment configuration—such as dual-cord servers, single-cord devices, liquid cooling, or a specified rack density—falls within the covered assumptions.
Specify measurement and calculation
Define where availability is measured, which monitoring system supplies the record, the calculation period, and how simultaneous or partial outages are counted. Require a process for obtaining incident records and disputing an incorrect classification. Clarify whether an event affecting one of your power paths but not the other is treated as an outage for your service.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallAddress maintenance and exclusions
List planned-maintenance notice requirements, allowed maintenance windows, emergency-work provisions, utility exclusions, tenant-caused events, force-majeure language, and exclusions for equipment outside the agreed load or environmental envelope. Make sure the exclusions do not silently remove the very power or cooling conditions your deployment requires.
Set notification, restoration, and remedies
Specify how quickly the operator must notify you, what information an incident notice must contain, and how restoration progress is communicated. Record restoration expectations where the operator can control them. Define service credits, any termination or step-down rights for repeated failures, and the procedure for claiming a remedy. There is no universal SLA target, credit level, generator runtime, or outage remedy; those terms depend on the facility and negotiation.
Compare two facilities on the same evidence
Use a common scorecard and mark unsupported claims as open items rather than awarding credit for vague assurances.
| Comparison axis | Evidence to collect | Decision question |
|---|---|---|
| Committed IT capacity | Reserved kW, measurement point, growth process, overage terms | Can the site support the initial and planned load without an unplanned commercial or technical change? |
| Power delivery path | One-line diagram, utility feeds, UPS, generators, transfer gear, final circuits | What serves the exact rack or room, and which elements are shared? |
| Maintenance behavior | Switching procedures, maintenance scenarios, test records, notices | Can required work occur without exceeding your risk tolerance? |
| Cooling envelope | Inlet limits, humidity, airflow, density limits, reserved capacity | Do actual equipment requirements remain inside the contractual environmental range? |
| High-density or liquid support | Interface specifications, supply conditions, controls, fault procedure | Is the proposed cooling method supported throughout normal operation and maintenance? |
| Redundancy evidence | Definitions of N+1 or 2N, path diagrams, common-point analysis | Does the stated redundancy survive a component failure and planned maintenance? |
| Certification scope | Current Tier or other validation documents and boundaries | Does the document apply to the space and service you are leasing? |
| SLA and remedies | Measurement, exclusions, notice, restoration, credits, termination rights | Can you prove a breach and obtain a meaningful remedy? |
| Total lease cost | Space, reserved power, metered power, cooling, upgrades, testing, and exit fees | What will the committed design cost at both initial and target load? |
Conduct a final diligence visit
Bring your electrical or mechanical engineer when the load, density, or risk justifies it. An independent data-center engineering review can help interpret diagrams, capacity studies, commissioning records, and maintenance evidence; select the reviewer based on relevant qualifications and independence rather than an operator referral.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsUse a document-and-observation checklist
- Match the one-line diagram to visible equipment labels and the proposed room or cage.
- Confirm that the rack positions, floor loading, containment, and cable routes in the proposal exist in the stated configuration.
- Ask to see alarm panels, monitoring screens, and the escalation roster used after hours.
- Review recent UPS, generator, transfer, and cooling-maintenance records, including exceptions and corrective actions.
- Confirm that construction, phased expansion, or nearby tenant work cannot remove a path on which your service depends.
- Record every open item with an owner, due date, and lease-document reference.
Recognize unresolved red flags
- The operator will not provide a current diagram or limits the discussion to a facility-wide Tier label.
- N+1 is stated without identifying the path, common components, or maintenance condition.
- Cooling capacity is quoted per room without a reserved density, inlet limit, or high-density policy.
- Generator runtime, fuel responsibility, or transfer behavior is described as “standard” but not documented.
- The SLA excludes utility, UPS, generator, or cooling events broadly enough to remove practical protection.
- Metering, overage rules, or capacity upgrades are left to a future order form.
Make the lease decision
Proceed only when three records agree: your load and environmental requirements, the facility’s demonstrated topology and operating procedures, and the lease’s measurable commitments. If one record is missing, classify the issue as a condition of contracting rather than an assumption.
For a lower-risk deployment, documented capacity, acceptable environmental limits, tested procedures, and clear incident remedies may be sufficient without the most elaborate topology. For life-critical, revenue-critical, or unusually dense workloads, require stronger evidence of independent paths, concurrent maintainability, maintenance-state performance, and recovery responsibilities. Local electrical and building codes, utility conditions, actual equipment loads, current operator documents, and legal review remain specific to the candidate site.
ASHRAE’s 2019 Chapter 20: Data Centers and Telecommunication Facilities puts the central lesson plainly: “However, the power and cooling redundancies can be significantly different from one colocation center to another, and should be evaluated before signing a contract, which should include a carefully worded SLA.” Treat that as a process requirement: inspect the exact service, test the operating story, and contract what matters.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
