Class C downlinks not transmitted to device after 2 hours

ChirpStack (v4)
1

I’d like to investigate the following problem, context is chirpstack 4.10.1 ; EU868 device in Class C ; connected on helium network:

The device joined and send an uplink, then I send downlinks to the device and it gets them. I stop any communication to the device for 2 hours (about) and then I send new downlink messages. They are now stored in queue and not propagated to the device.

A - Is this behavior normal ?
B - what prevent Chirpstack to push the message to the gateway after a certain time ?
C - is there a parameter to extend this kind of timeout ?

2

Did you specify that the device is class C?

Is the problem repeatable? I am still at 4.6.0 for my production system. Sometimes, tbe device joins, but Chirpstack fails to prepare its internal data structures correctly and downlinks cannot be scheduled by the class C scheduler until the device re-joins.

3

Device is configured class C, device profile is class C also, It works during the first hour, stop after. This is repeatable.