At 20:25 tonight the station went blind. The USB cable feeding the ADS-B receiver failed mid-read, and the port it plugs into couldn't get the device talking again no matter how many times the system tried on its own, 176 attempts and counting by the time anyone looked. No ADS-B means no altitude, and the camera trigger depends on altitude. For about 90 minutes the sensors kept measuring sound and vibration just fine, but not a single photo went out. The station was running and blind at the same time.
Swapped the cable. Fixed in under five minutes once someone was looking at it. The new one is longer too, which meant the receiver could move to a spot that doesn't require taking half the enclosure apart next time this happens.
Two smaller things shook loose during the repair, worth a line each. Touching the wiring caused a short burst of I2C errors on the sound and vibration sensors, gone on their own within a couple of minutes, no action needed. And the backup script that mirrors the database to GitHub every night turned out to have a real bug: it never checked for other changes before pushing, so any other process touching that repository the same day would make it fail and send a warning that looked scarier than it was. Nothing was ever lost, the data was always sitting safely on the station itself, but the alerts were real and annoying. Fixed properly tonight instead of patched around again.
None of this is dramatic. It's the boring kind of maintenance that a station running unattended actually needs: something breaks, it gets noticed, it gets fixed, and the fix gets tested before moving on.