Loading
Preparing page…
Checklist
Use this when you inherit a server room, prepare a refresh, or need an honest baseline before change work. The goal is a verified picture of what is installed—not a tidy folder that no longer matches the floor.
Server rooms accumulate drift. Racks gain devices without elevation updates, PDU outlets are reused without circuit notes, temporary cables become permanent, and labels fade or never matched a standard. Monitoring may cover hosts while ignoring environmental or power signals that predict outages. An audit closes the gap between assumed design and operable truth.
This checklist is for on-site and document review together. Walk the room with the latest drawings and asset lists in hand. Where they disagree, the room wins until documentation is corrected. The output should be a punch list with owners and priorities—not a narrative that cannot drive remediation.
Treat the audit as an engineering input for capacity planning, cabling remediation, access hardening, and monitoring coverage. Without a verified baseline, any automation or expansion work will encode the same errors into the next phase of delivery.
Prepare by collecting rack elevations, circuit maps, cable schedules, IP plans, access lists, and monitoring inventories. Mark what is missing before you walk the floor so field time focuses on verification, not discovery of which documents exist.
Work rack by rack and circuit by circuit. Photograph labels and panel faces where helpful, but record structured findings: device, location, discrepancy, risk, and recommended action. Separate immediate safety or access issues from documentation debt that can wait for a planned window.
Close the audit with an updated as-built package or an explicit backlog to produce one. An audit that only produces slides does not improve the next incident response.
For each rack, verify device identity, rack unit position, orientation, and blanking. Record undocumented devices and empty U that drawings still show as occupied.
Match physical tags to CMDB or spreadsheet inventory. Flag devices with no tag, duplicate tags, or serial numbers that do not match records.
From console or management interfaces, confirm hostnames and management IPs match documentation. Note orphaned addresses and devices reachable only by tribal knowledge.
Identify dead or spare equipment still consuming power, space, or network ports. Decide remove, quarantine, or document as intentional spare.
Identify which drawing set is authoritative, who last updated it, and when. If multiple conflicting versions circulate, designate one source of truth and archive the rest.
Capture floor-standing gear, temporary switches, and trial hardware. Either schedule removal or promote them into the formal design with power and cable impact assessed.
Trace or label-read from device power cord to PDU outlet to breaker or feed. Gaps here are the most common cause of accidental outages during maintenance.
Record current load per PDU or circuit where meters exist. Flag circuits near capacity and any single-corded critical devices on one feed when dual feeds were intended.
Confirm A/B or primary/secondary labeling matches how devices are plugged. Test understanding with operators: can they state which feed loss takes which services down?
Look for power strips in series, undersized cords, or shared outlets that bypass the designed PDU plan. Treat these as priority remediation.
Observe intake/exhaust orientation, blanking, and obstruction of perforated tiles or vents. Note hot spots and whether sensors (if any) sit where they represent rack intake.
If sensors exist, verify they alert somewhere operators watch. If none exist in a dense room, record the gap as a monitoring finding, not an aesthetic note.
Confirm which racks or circuits sit on UPS, runtime expectations, and who owns test schedules. Unverified runtime claims are planning risks.
Note overloaded trays, blocked airflow from cable bundles, and runs that violate bend radius or firestop. Photograph before/after candidates for remediation windows.
Pick critical paths (uplinks, storage, management) and confirm labels at both ends match documentation. Sample enough links to estimate systemic labeling debt.
Compare panel schedules to live patching. Record abandoned patches still connected and live links with no schedule entry.
Flag cables with no label or no schedule row. Either label and document them or schedule removal after proving they are unused.
If a standard exists (e.g. management vs production colours), score compliance. Inconsistent conventions increase change risk more than missing aesthetics.
Look for cut ends, abandoned trunks, and debris that complicate future installs. Add cleanup to the punch list where it blocks pathways.
Check that matching patch lengths and optics are available for common failure replacements. Missing spares turn simple faults into extended outages.
Compare badge or key holders to who actually enters. Note propped doors, shared credentials, and visitor processes that are undocumented.
From the intended management path, confirm KVM, serial, and BMC access for critical devices. Record devices that require physical presence only.
If security controls are stated in policy, verify they work and that logs are retained somewhere reviewable. Paper policy without working controls is a finding.
List devices and links with no health checks, and checks that point at decommissioned targets. Prefer coverage of power, environment, uplinks, and critical services first.
Trigger or review recent alerts to confirm they reach people who can act. Dead email lists and abandoned chat channels count as monitoring gaps.
Confirm authentication and config-change logs for switches and jump hosts land in a durable store. Note retention and who can query them during an incident.
Group findings into safety/power, access, cable/label, documentation, and monitoring. Assign owners and target windows. The audit is incomplete without this list.
Plan a shorter re-walk for high-priority remediations. Closing tickets without re-checking the floor recreates the same drift.
Services
Evidence
Industries
Resources
Next step
Share scope and constraints. We reply within 1–2 business days with fit and a practical approach.
AstraMakers