Video of Every Machine Stoppage, Captured Without Anyone Pressing Record

What happens between the stop and the clip

1. The call is raised. An operator, or an EdgeSense device at the machine, places a downtime call. Creating the call is the trigger; there is no separate recording step.
2. The lead-up is retrieved. Each camera records continuously into a rolling buffer whose length is set per camera, in minutes. The clip takes the window before the trigger from that buffer.
3. Recording continues. Capture runs for a configurable number of seconds after the trigger, so the machine’s response and what the operator did next are kept too.
4. The clip is linked. A background service processes the clip, generates a thumbnail, links it to the downtime call in the database and pushes a “video ready” notice to the people involved.
5. It is on the work order. The technician reviewing the work order sees a Video Evidence section with the clip. No recorder login and no timestamp search.

In the example on our video integration page, a press clip kept 45 seconds before the call and 15 after: sixty seconds of evidence nobody had to ask for.

Why plants rarely have video of their failures

Most plants already have cameras. What they rarely have is the right minute of footage, kept, and attached to the failure it shows.

• The footage is overwritten. A network video recorder keeps a rolling window of days or weeks. By the time a recurring failure is recognized, the first occurrence is gone.
• Finding the clip is a project. The right camera, the right timestamp and an export usually need someone from IT, hours after the context has faded.
• The video and the event live apart. Matching a clip to a downtime event, a work order or a scrap entry means comparing timestamps across two systems by hand.

Capturing on the event and storing the clip with the record removes all three. The streams stay separate, video in its own stream and machine data in its own, but they land in one database on the same timeline, so they can be read against each other with nothing to reconcile.

Cameras, storage and review

• Cameras. Any IP camera that exposes a standard RTSP stream, for example Reolink, Hikvision, Dahua or Axis. More than one camera can be assigned to an asset, and the player switches between them.
• Storage. As a typical figure, H.264 at 2 Mbps for 1080p gives about 75 MB for a five-minute clip, or roughly 90 GB a year for a facility with 1,200 downtime calls. Standard NAS (Network Attached Storage) or server storage is enough for most plants.
• Review. Frame-by-frame stepping, playback from 0.2x to 3x and digital zoom, built for failure analysis rather than surveillance.
• Reliability. A clip that produced no footage never shows as ready, and the failure is reported to the people who can act on it. A recorder that stalls detects it and recovers. Deleting a camera keeps the clips already recorded against it.

Questions buyers ask next

Can software capture video of a machine downtime event automatically?

Yes. In ProAlert, raising a downtime call on an asset triggers a clip from each RTSP IP camera assigned to that asset. The clip covers a lead-up taken from the camera rolling buffer and a configurable number of seconds after the trigger, and it is linked to the downtime call and shown on the resulting work order. Nobody has to press record.

Does the clip include what happened before the machine stopped?

Yes. Each camera records continuously into a rolling buffer whose length is set per camera in minutes, and the clip takes the window before the trigger from that buffer, so the cause is in the frame rather than only the aftermath.

Do we need special cameras or a separate video management system?

No proprietary cameras are required. Any IP camera with a standard RTSP stream works, and the video management module runs inside ProAlert on the same database as maintenance and OEE. Video is a licensed module.

How much storage does downtime video need?

As a typical figure, H.264 at 2 Mbps for 1080p produces about 75 MB for a five-minute clip, which is roughly 90 GB a year for a facility with 1,200 downtime calls. Standard NAS or server storage is sufficient for most deployments.