We are willing to start using watchdog sample in our CHESTER applications.
Is there a general guideline to follow?
How can we fine tune the watchdog timer according to our different type of applications.
We appreciate any help.
in our catalogue application, we use the basic functionality, enable watchdog for generous 120 seconds by
ctr_wdog_set_timeout(120000); and feed it in the main task, which blinks the LED by calling
Since the device is not reacting to some real-time events and it just sends data every 10 minutes, this is a good compromise.
Our SDK does not use any busy-waiting, so the risk that code gets frozen in some infinite loop is basically zero.
In some projects, we have a global 32-bit variable where we enable bits from different running tasks. Then in the main or dedicated task, we check that variable, and if it has set correct bits from all the tasks, we are sure that all the tasks are running, and only then we feed the watchdog. This way the watchdog triggers even if a single task stops.