Engineering guide

What ‘last seen’ really means in endpoint monitoring

A device status indicator is only as trustworthy as the heartbeat model behind it.

“Online” sounds simple, but it is a derived state. Most endpoint platforms infer it from the most recent successful communication between an agent and a control plane.

Heartbeat interval

If an agent checks in every 30 seconds, a platform should not normally mark it offline after 31 seconds. Networks jitter, laptops sleep and processes can briefly pause.

Status threshold

A useful threshold creates tolerance. For example, a 30-second heartbeat might use a 90-second online threshold. That gives the device several opportunities to report before the UI changes state.

Stale versus offline

At larger scale it can be useful to distinguish online, stale and offline. A stale state tells an operator that evidence is getting old without immediately declaring failure.

Store timestamps consistently

Store server timestamps in a consistent standard such as UTC and convert them for the viewer in the browser. That prevents multi-country fleets from mixing storage logic with presentation logic.

Why this matters

Operational decisions should not be driven by an attractive green dot unless the platform can explain exactly how that dot was calculated.

Try the synthetic status model.

The ZENQIX demo lets you filter online and offline endpoints without exposing any real customer data.

Open Demo Hub