Fixed watchdog not logging, increased cycle trigger

[alexcekay]

Solved Problem

Looking through some logs, the following problems were found:

• During MAVLINK logging startup, the logger requires a lot of communication (including ACKs) with the companion. This burst can sometimes cause the logger's main cycle to take more than 3 seconds, triggering the watchdog.
• If a load measurement is already active, a watchdog trigger will not be detected.
• The watchdog does not indicate whether it was caused by the logger being in the ready state for too long or by the logger main cycle taking too long.

Solution

• The cycle timeout has been increased from 3s to 5s.
• If a load measurement is already active and the watchdog triggers, the cause of the load measurement is overwritten with "Watchdog".
• The watchdog error log contains the reason for the trigger.

Changelog Entry

For release notes:

Bugfix Fixed watchdog not logging and increased cycle timeout

Test coverage

• Tested on the bench with a Skynode v5x

[bkueng]

Nice, makes sense