Anatomy of a gap
Gaps come in three shapes. Hard gaps: recording stopped entirely for a period - a crashed service, a full disk, a rebooting recorder. Soft gaps: recording continued but intermittently - an unstable camera stream dropping in and out, motion detection missing events. Retention gaps: the footage existed and was overwritten earlier than expected, so the gap is at the old end of the archive rather than the new end.
All three are invisible to live monitoring, and the third is invisible even to spot playback checks - yesterday plays fine; the missing footage is from three weeks ago, and nobody looks there until it matters.
Why gaps cluster where it hurts
Power and network events
Gaps correlate with site power dips and network outages - the same events that correlate with break-ins and disputes. The minutes most likely to be requested are the minutes most likely to be missing.
Nights and weekends
Storage pressure peaks and nobody is watching. Failures that begin Friday evening run unobserved until Monday - a 60-hour gap from a fault that would have been a 2-hour gap on a Tuesday.
The busiest cameras
High-traffic scenes produce the most motion events and the highest bitrates, hitting storage and encoder limits first - so entrance and dock cameras gap before the corridor nobody films.
The real cost of a gapped archive
Operationally, a gap converts an answerable question - what happened at the loading dock at 02:40 - into a permanent unknown, with everything that follows: unresolvable disputes, insurance claims paid without evidence, internal investigations that end in suspicion instead of facts.
Where retention is contractually or legally required, a gapped archive is also a compliance exposure: the requirement is rarely 'record most of the time'. And in legal contexts, systematic gaps invite the worst interpretation - opposing counsel does not assume your disk was full by accident.
How continuous gap detection works
- 01
Segment inventory per camera
On every check cycle, the recording system is queried for its actual segments per channel - the same information playback uses, read programmatically.
- 02
Freshness and continuity evaluation
The newest segment must be recent, and the sequence must be continuous within the channel's expected recording pattern. A missing stretch becomes a measured gap with start time and duration.
- 03
Incident with evidence
The gap opens an incident for the exact camera, correlated with storage, clock and device events to point at the root cause - and routed as a ticket to whoever owns the fix.
Measured gaps change the conversation
Once gaps are measured instead of suspected, the operational conversation changes. 'Camera 12 seems unreliable' becomes 'camera 12 gapped four times this month, always during the 22:00 backup window' - which is a network QoS problem with an obvious fix, not a camera problem. Gap history per camera and site is also the honest metric for maintenance vendors: not visits performed, but continuity delivered.
This is the model IOSentra implements: scheduled per-camera segment verification, measured gap records, root-cause correlation and SLA-routed resolution - so a gap is a maintenance event that lasts hours, not a discovery that ends an investigation.