The maintenance report says “all devices are working”, but on the night of an incident no face can be made out from the entrance gate camera. Periodic maintenance answers “are the devices working?”; a technical audit answers, with evidence, “will this system give useful footage at the moment of an incident, is it secure, is it sustainable?”. Below, for IT and security managers, we explain the audit scope, sampling method, evidence collection practices and the structure of an actionable findings report.
When a technical audit is needed
Triggering events are as decisive as the calendar. If any of the following applies, plan an independent technical assessment:
- A recording searched for after an incident couldn’t be found, or the footage wasn’t enough to identify a person or plate.
- A change of maintenance company or managed service provider; recording the current state before handover.
- Establishing the baseline before a major renewal, VMS migration or capacity increase.
- A change in building use: a new production line, warehouse layout or entry points.
- Technical verification of physical security controls before an ISO 27001, TISAX or customer audit.
- Situations where the system has grown undocumented and it’s unknown who manages what.
Compliance-focused controls (access permissions, retention, privacy notices) are covered in the TISAX and ISO 27001 camera controls article. A technical audit measures the technical reality beneath those controls. Policy is on paper; reality is on site.
Audit scope and areas
The scope is agreed in writing before the audit starts. For a holistic assessment of camera systems, six areas are enough:
| Field | What is assessed | Typical evidence |
|---|---|---|
| Image quality | Reaching the target image level (DORI), night performance, WDR | Day/night snapshots, pixel density measurement |
| Coverage | Seeing critical points, blind spots, framing drift | Coverage marked on the floor plan, reference comparison |
| Record | Recording mode, retention period, gaps, clock sync | VMS timeline, retention report, NTP status |
| Network | Segmentation, PoE budget, uplink usage, port errors | Switch outputs, VLAN and ACL configuration |
| Security | Default accounts, firmware, open services, remote access | Port scan, account list, firmware version table |
| Power and environment | UPS runtime, cabinet temperature, grounding, outdoor protection | UPS test log, temperature measurement, site photos |
What’s excluded should also be written down. Legal compliance assessment, structural (pole, façade) inspection or penetration testing require separate expertise and are noted in the report as “out of scope”. An out-of-scope item that isn’t written down later turns into the argument “you should have looked at this too”.
Sampling and site method
In a system with a few hundred cameras, inspecting every camera on site one by one usually isn’t necessary. The right method is to apply remotely feasible checks to the whole population, and checks that must be done on site to a sample.
Checks applied to the whole population
- Firmware versions, model distribution and support status (via inventory and ONVIF queries).
- Recording status, retention period and recording gaps over the last 30 days (via VMS reports).
- Clock sync: all cameras and recording servers tied to the same NTP source.
- A day and night snapshot from each camera (taken remotely, with a preliminary assessment of framing and quality).
The item on this list that produces the most surprises is clock sync. A camera that has drifted a few minutes breaks the sequence of footage from different cameras in an incident review.
Site checks applied by sampling
The site sample is chosen on a risk basis, not at random. The approach we recommend: all cameras at critical points (entrances, checkouts, shipping, perimeter); all cameras where a problem was seen in the remote preliminary assessment; and from the remaining population, a sample in which every model, every building and every mounting type (indoor, outdoor, high pole) is represented. Write in the report how the sample was chosen and its size, with the reasoning; only then can the reader understand to what extent findings can be generalized to the system as a whole. For the general principles of sampling, the sample selection in internal audit article is complementary.
For image quality, the method is comparison with the image level the system was designed for. Target levels and the pixel density calculation are explained in the IP camera system design guide; the standard on which the image level classification in video surveillance systems is based is IEC 62676-4 (IEC). On site, a measurement is taken at the relevant distance with a target of known size or a test chart, and repeated both by day and at night. A daytime test alone isn’t enough. Most identification problems appear at night.
Evidence collection: screenshots, logs, measurements
A finding not supported by evidence is no more than an opinion. Every piece of evidence is recorded with its source, timestamp and the relevant device ID.
- Screenshots: VMS configuration pages, recording schedule, user list. The system time and device name should be visible in the screenshot.
- Snapshots: From the camera itself or the VMS, with the camera name and timestamp.
- Log and report exports: Recording gap report, event logs, user session logs.
- Configuration outputs: Switch, VLAN, ACL and PoE status.
- Measurements: UPS runtime test, cabinet temperature, cable test results.
- Site photos: Mounting, cable entry, cabinet layout; with location information.
Some commands that can be used in the network and security area (Cisco IOS and Linux example):
# PoE budget and per-port power status (Cisco IOS)
show power inline
# Error counters on ports in the camera VLAN
show interfaces counters errors
# Identifying open management and streaming services in the camera VLAN (authorized scan)
nmap -sV -p 21,22,23,80,443,554,8000,8080 10.20.0.0/24 -oA camera-vlan-scan
Run the scan only with written permission and in a time window that won’t affect the recording system. Some old cameras reboot under heavy scanning; try a small subnet first. The auditor’s scan mustn’t produce recording gaps.
Findings classification table
Findings are classified on a consistent scale by impact and likelihood. The four-level scale below is a practical framework for camera systems:
| Level | Definition | Example finding | Recommended timeframe |
|---|---|---|---|
| Critical | Recording loss or unauthorized access is present or imminent | NVR management interface exposed to the internet; a critical camera not recording | Immediately |
| High | Evidence can’t be produced at the moment of an incident, or a serious security vulnerability | Entrance camera doesn’t reach identification level at night; default password | 30 days |
| Intermediate | A condition that reduces the system’s reliability or manageability | No clock sync; out-of-support firmware | 90 days |
| Low / observation | An improvement opportunity | Inconsistent naming; missing cable labels | During planned maintenance |
The timeframes are recommendations; they’re set together with management according to the organization’s risk appetite. The real point is that the same finding isn’t classified at different levels in two different audits. For that, each level should have a written guide with examples.
Structure of the audit report
The report is layered according to its reader: an executive should be able to decide in two pages, and the technical team should be able to reproduce every finding. Recommended section structure:
- Executive summary: Overall assessment, number of findings by level, the top three priority actions.
- Scope and method: Sites and systems audited, out-of-scope areas, sampling method and size, audit dates.
- Results by area: A short assessment for each of the six areas.
- Finding records: For each finding, an ID number, title, level, affected devices, criterion (against what it falls short), current state, risk, evidence reference and recommendation.
- Appendices: Evidence files, device list, snapshot archive, measurement tables.
When a finding is written with a “criterion–condition–cause–effect” structure, little room is left for dispute: “The entrance gate camera is expected to provide identification level at night (criterion); in the night image, pixel density is below target and facial detail can’t be made out (condition); the IR range isn’t enough for the distance and WDR is off (cause); people can’t be identified in night-time entry events (effect).” A common mistake is writing only “camera image insufficient” and leaving it there. If the cause isn’t written, the fix is guesswork too.
Tracking corrective actions
The report’s value emerges when findings are closed. The structure we recommend for tracking:
- For each finding, the responsible person, target date and accepted fix method are written down.
- For findings that can’t be fixed, an authorized manager signs a risk acceptance and the reason is recorded.
- Closure is verified with new evidence: a new night image from the fixed camera, the audit log of the changed password, a screenshot of the new firmware version.
- An interim check is done for critical and high findings, and verification for all findings at the next audit.
- Recurring findings are examined for root cause; if the same problem appears in every audit, the process is lacking.
Some fixes turn into routine work and are added to the periodic maintenance checklist. An up-to-date inventory is essential for the audit to be done quickly; the inventory structure is explained in the camera inventory guide. For the scope of an independent assessment, see the camera systems technical audit page, and for redesign needs after findings, the survey and consulting page.
Frequently asked questions
What’s the difference between a camera system audit and periodic maintenance?
Periodic maintenance is routine work to keep devices running. A technical audit assesses, based on evidence, whether the system meets its purpose in terms of image quality, coverage, recording, network, security and power, and produces risk-classified findings.
Should every camera be inspected on site in an audit?
No. Firmware, recording, clock sync and snapshot checks are applied remotely to all cameras. Site checks cover all critical points, cameras that appear problematic and a risk-based sample representing each model and mounting type.
Should a critical gap found during the audit be fixed immediately?
The auditor doesn’t change the configuration. When a critical risk is seen, the system owner is notified in writing immediately, the system owner carries out the fix, and the result is documented in the report with new evidence.
How often should a camera system be audited?
There’s no fixed mandatory frequency. In practice, a comprehensive assessment once a year is recommended, plus additional audits on triggers such as a change of service provider, a major renewal or an incident where the recording couldn’t be found.