The picture is still there: handling a camera that stopped updating

By Gate & Shift Editorial Desk · Planned week of 2026-11-30 · Documentation reviewed 2026-11-15

A still image can look perfectly ordinary at a quiet gate. The operational question is whether it is current, so put a timestamp check into the shift routine — particularly before somebody signs “no activity” on the overnight sheet.

This guide proposes a local operating procedure. It makes no claim that every product displays health information in the same place or sends an offline notification.

Establish the last time you can support

Write down the image timestamp, the checking device's current time and any application health indicator. Keep the screen if your evidence policy allows it. If the two clocks disagree, record the discrepancy without silently correcting either time in the log.

Mark where the known gap begins. If all you know is when staff noticed it, label the earlier boundary uncertain; a confident-looking start time is not worth inventing. Coverage calculations need a declared schedule and a record of unobserved intervals under our house method M02.

Draw the path the picture takes

Draw four boxes on paper: camera, local recorder or connector, network, viewing/analytics service. Then ask the installer which component needs the internet for each function. “Cloud system” is too broad to answer whether a local recording exists to recover.

For example, Coram documents local storage on Point V03, and Lumana describes Core as a replacement recorder storing and processing footage V06. Eagle Eye describes Bridge/CMVR connections for most installations, with Camera Direct available for eligible smaller ones V11. Those architectures justify different questions. They do not, on their own, establish how every alert or remote view behaves during your outage.

Put a person on the gap

Name who checks the network, who contacts the installer and who authorizes a temporary alternative. A night supervisor should not be making an unapproved firewall change from an emailed suggestion; use the organization's normal support process and record its reference number.

Keep the gap on the handover even if the picture returns before morning. The incoming crew needs to know what could not be observed, not just the current green status.

Verify recovery with movement and a timestamp

Once the feed returns, confirm that the timestamp advances and the view is the intended one. Use an approved visible movement or a permitted test recording. If the alert workflow is in scope, check it separately — a recovered live picture does not establish that the recipient route has recovered too.

Record recovery time, test result and any changed setting; the shift handover sheet has places for each. Keep the initial failed check rather than replacing it with a successful screenshot.

Calculate what the shift actually covered

Here is a hypothetical example. A camera was scheduled for eight hours and was demonstrably unavailable for ninety minutes. Its observed coverage is 6.5/8 = 81.25%, assuming the rest of the scheduled interval was monitored. Don't describe that as a complete uneventful watch.

Our proposed method reports coverage separately from recall and alert counts M02. If the outage start is uncertain, say so beside the fraction. A range or an unresolved gap tells a buyer more than a precise percentage built on a guessed timestamp.

FAQ

Can the incident list tell me whether the feed was up?

An empty incident list alone cannot establish observation coverage. Review whatever separate health or coverage evidence the actual system provides.

Should we count an outage as zero false alerts?

Keep unmonitored time out of a monitored-time nuisance-alert denominator and report the gap separately. Otherwise a disconnected camera can appear to have excellent alert performance.