Spinning responses from modem should not lock out usage

[JimCircadian]

Switching the modem on at the wrong time (as these days it's not necessarily power controlled by us) can lead to an endless loop in response handling

12:30:29                 DEBUG            Next wakeup is due at 2025-12-01 23:01:00+00:00 (in 37830.802159 seconds)
12:30:29                 INFO             Job "DummyTask (trigger: date[2025-12-01 12:30:12 UTC], next run at: 2025-12-01 12:30:12 UTC)" executed successfully
12:30:30                 INFO             Creating pyserial comms instance to modem
12:30:30                 INFO             Command sent: "AT+CGMM"
12:30:31                 DEBUG            Reply received: '... Version: ... Serial Number: ...'
12:30:31                 DEBUG            Reply received: '... Version: ... Serial Number: ...'
12:30:31                 DEBUG            Reply received: '... Version: ... Serial Number: ...'
12:30:31                 DEBUG            Reply received: '... Version: ... Serial Number: ...'
12:30:31                 DEBUG            Reply received: '... Version: ... Serial Number: ...'
12:30:31                 DEBUG            Reply received: '... Version: ... Serial Number: ...'
12:30:31                 DEBUG            Reply received: '... Version: ... Serial Number: ...'
12:30:31                 DEBUG            Reply received: '... Version: ... Serial Number: ...'

[JimCircadian]

This is a fringe event for our installations, but needs defense against!