CDFOM cheat sheet

MTTR = total repair time ÷ number of repairs; MTBF = total operating time ÷ number of failures. Availability from these = MTBF ÷ (MTBF + MTTR).

SLA tiers are typically expressed as monthly uptime: 99.9% allows ~43.2 min downtime/month, 99.99% ~4.32 min, 99.999% ~26 s; annual equivalents are 8.76 h, 52.6 min and 5.26 min.

Incident vs problem: an incident is an unplanned interruption or degradation restored by a workaround; a problem is the underlying cause, closed only by a permanent fix or accepted risk.

Change categories: standard (pre-approved, low risk), normal (CAB review), emergency (expedited approval, retrospective review within an agreed window).

Escalation has two axes: functional (to deeper expertise) and hierarchical (to higher authority); both need defined trigger times in the operations manual.

KPI set commonly includes PUE, SLA attainment %, incident count by severity, MTTR, change success rate, first-time-fix rate and capacity headroom %.

PUE = total facility energy ÷ IT equipment energy; it is a ratio with no unit and is measured over a defined period, not instantaneously.

Shift handover record must state open incidents, pending changes, active permits, alarms suppressed or inhibited, and outstanding vendor visits.

Planned maintenance splits into preventive (time- or usage-based) and predictive (condition-based, triggered by trend data such as vibration or thermal rise).

Reactive maintenance is corrective after failure; deferred corrective work must be risk-ranked, not simply queued by date.

Maintenance contract types: comprehensive (parts, labour, consumables included), non-comprehensive (labour only), and OEM service level agreements tied to warranty terms.

Key contract terms to verify: response time vs fix time, escalation path, spares held on site, reporting obligations, KPI credits, and exclusions.

Battery strings: IEEE 1188-style practice uses periodic impedance/conductance checks; a rising impedance trend against baseline flags a weak jar before capacity test failure.

UPS and generator testing cadence commonly: UPS battery test monthly, full load bank annually; generator exercise monthly with annual load-bank test at rated capacity.

Infrared thermography of electrical distribution should be done under load; a loaded scan is required for meaningful results, and findings are ranked by temperature delta.

CMMS records must link asset, work order, procedure, technician, spares consumed and outcome, so that maintenance history supports reliability analysis.

Electrical arc flash boundary and incident energy depend on available fault current, clearing time and working distance; work inside the boundary requires rated PPE and an energised work permit.

Lockout/tagout sequence: notify, shut down, isolate, lock and tag, release stored energy, verify absence of voltage, then work; removal reverses with verification.

NFPA 75/76-style guidance: detection should be cross-zoned (two independent detectors) before agent release or power isolation, to limit false discharges.

Gaseous suppression agents (clean agents) require room integrity testing; a door fan test verifies the enclosure holds agent concentration for the designed hold time.

Environmental envelope is commonly 18-27 °C and 40-60% relative humidity for ASHRAE A1-class equipment; excursions must be logged and correlated with hardware faults.

Physical security layers: perimeter, building envelope, data hall, cabinet, and logical access; each layer needs its own access control, logging and alarm response.

Visitor management: pre-registration, identity verification, escorted access, badge issue and return, and retention of access logs for the audit period.

High-availability fire detection must not be disabled for hot work without a documented impairment permit and compensatory measures such as fire watch.

Shift coverage formula: required FTEs = (24 h × 7 d × positions staffed) ÷ (average weekly hours per FTE × availability factor), then add relief for leave and sickness.

A common relief factor is 1.15-1.25 for planned absence and training; critical sites often staff to N+1 per shift so one absence does not reduce below minimum.

Competence is evidenced by three elements: training completed, assessment passed, and supervised practice signed off by an authorised assessor.

Authorisation to perform switching or work on critical systems should be a named, time-limited, revocable permit held by the individual, not a team attribute.

Shift pattern design must respect local working-time rules on maximum consecutive hours, minimum rest between shifts, and night-work limits.

Skills matrix maps each role against required competencies; gaps are shown as a dated training plan with an owner, not as a narrative.

Vendor and contractor staff require site induction, escort or authorisation level, and competence evidence before unescorted access to critical areas.

Risk score = likelihood × impact; each risk register entry needs an owner, a treatment (avoid, reduce, transfer, accept), a target date and a residual score.

ISO/IEC 27001 covers information security management; ISO 22301 covers business continuity; ISO 50001 covers energy management; ISO 41001 covers facility management.

An audit finding is classified by severity (major, minor, observation); major findings require root-cause analysis and a corrective action plan with a verified closure date.

Document control requires version number, effective date, author, approver, and a defined review cycle; superseded documents must be withdrawn from use.

Separation of duties: the person who approves a change should not be the sole person who implements and verifies it on critical systems.

A risk register differs from an issue log: risks are uncertain future events, issues are realised events requiring action now.

Evidence retention for compliance is typically aligned to the longest applicable requirement among contract, certification body and regulator.

Concurrent maintainability is the property that any single component can be maintained while the system remains available; it requires N+1 or better at every dependency, including power and cooling paths.

Tier-style topology: N = no redundancy, N+1 = one redundant component, 2N = fully duplicated paths, 2N+1 = duplicated plus one extra; verify the actual path, not the nameplate.

Capacity headroom is tracked as % of design capacity used; thresholds commonly trigger action at 70% (plan), 80% (order), 90% (expedite).

Power capacity check: rack kW × quantity plus cooling and losses, compared against UPS and PDU ratings with derating applied for continuous load.

RTO is the maximum acceptable time to restore a service; RPO is the maximum acceptable data loss measured in time. They drive backup and replication design.

Single points of failure are found by tracing each utility path end to end, including control systems, and confirming no shared component exists.

Continuity exercises should include a no-notice element and a documented debrief; untested plans are treated as unproven in audit.

Cooling capacity depends on delta-T and airflow as well as nominal kW; hot spots can occur well below nominal capacity due to bypass airflow.