The station disappeared from a monitoring room. Later it returned with its historical data intact. Stranger still, it could appear healthy in one monitoring room while remaining offline in another.
“Why is the station offline?” sounded reasonable. It was also hiding several different systems inside one word.
Offline where?
Separate the layers
Was the instrument measuring? Was the logger storing? Was the modem connected? Was the server receiving? Was one user interface displaying the latest data while another endpoint was not?
The returning history mattered. It suggested that loss of live visibility did not necessarily mean loss of local measurement. The receiving endpoints also mattered because a station can be reachable from one route and absent from another.
A better diagnostic question
Troubleshooting improved when “offline” became a location rather than a diagnosis. Measurement, storage, transport, ingestion and display could then be tested separately.
The lesson travels well: loss of telemetry is not automatically loss of measurement.