As here, a system with an ATmega328P using sleep modes, with a regular watchdog wakeup (for a timing 'tick').
The WD is used just for its interrupt - not for resetting.
But I find that the WDT period changes depending on whether I use Standby or Power Down sleep mode:
Standby current consumption:
The peaks are the (brief) wakeups on the WDT interrupts; the wakeup period is ~ 35ms
The wakeup period is ~ 41ms
The Datasheet says the WDT is a separate oscillator, and I can't see any mention of any dependence on or relation to the sleep mode - so why should the WDT change its period in the different sleep modes?
I have 4 boards, and they all have similar behaviour.
The AVR has an external crystal; the key difference between Standby and Power Down (AFAIKS) is that the oscillator stays running in Standby, but stops in Power Down; but why should that make any difference to the WDT - which is a separate oscillator?
(the current consumption traces are from the Microchip Data Visualiser, using the measurement module on a SAML21 Xplained-Pro)