resets,WDTs and hold

[kjm]

I've got a .hold(1) on gpy relay driver pin P23 so that if the program does a machine.reset() or the wdt=machine.WDT times out the relay state is preserved during the reboot. This works nicely and an ON relay is still on after the

rst:0x7 (TG0WDT_SYS_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT)
configsip: 0, SPIWP:0xee
clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00
mode:DIO, clock div:1
load:0x3fff8020,len:8
load:0x3fff8028,len:2136
load:0x4009fa00,len:19944
entry 0x400a05e8

event. Occasionally the GPY will reset for no apparent reason! There is no core panic style output, just the regular reset text above. These unexplained resets have the annoying effect that .hold(1) functionality is lost & an ON relay clicks off during the reboot. Wondering if there is some sort of trace I can run to maybe get a handle on what's going on?