Last night at 22:13 the station lost its Wi-Fi connection and did not reconnect on its own. For about ten hours, nothing reached the website: no live data, no new flights on the home page.
But the station never stopped working. The aircraft receiver, the microphone and the vibration and air quality sensors are all local, so it kept doing its job. 77 flights were recorded overnight, each with its noise peak and vibration, plus the continuous sound and vibration logs, second by second. Nothing was lost on the measurement side.
Two things did go missing. The camera also uses the Wi-Fi, so those 77 flights have no photo. And the weather comes from the internet: every flight overnight was tagged with the same reading from 22:13. Those values have been replaced with the real quarter-hourly weather for the night.
The first to notice was not me. Lisbon Finals also streams its measurements live, second by second, to SensorGrid, the sensor network run by Alessandro in Milan. Early this morning he wrote to ask why he could no longer hear us. When the station goes quiet, someone on the other side of the network notices. The missing hours are being sent to him now, from the station's own records.
A power restart at 08:05 brought everything back. One American Airlines 787 passed overhead seconds after a later service restart, too early to be measured live. Its true peak, 91.2 dB, was taken from the continuous log.
Two changes so it does not happen again: Wi-Fi power saving is now off, and the station now restarts its own connection, then reboots itself if needed, when the network stays down. The researcher data files have been refreshed with the corrected night.
Ten hours offline, and the measurements never stopped. Good test.