Ok, I’ll reply to my own message 
I took the unit apart today and poked around. I’m pretty sure the relay was fine, obviously the zwave module and tilt sensor seemed fine too based on accurate door status and always responding to ST requests with beeping.
I still found it interesting that the module would respond to successive open or close requests with no time delay - as if the module knew it had not opened or closed the door.
I think the module can sense the LED operation (voltage across led within some range, it seems) and mine may have failed. I presume the LS testpoint is the feedback path from the LED to the micro.
I pulled the LED off the board and did some measurements, the drive appeared to work ok, and pushed out 1 amp from what I believe to be a constant current source, enabled by the micro @ 3.3v on the testpoint labeled LD.
I built a diode and 2.4ohm resistor combination in series and placed it in place of the LED. Of course this will not light up, but should be roughly the LED forward drop with 1A flowing through it. I am guessing I could have done a single 3.5 ohm resistor instead (no diode), but I didn’t try it. I picked a 1n4004 and 3W 2.4 ohm resistor. Technically I’m over powering my diode and the resistor should be derated as well if 100% on, but given the pulsed output for a short time, I should be ok for now - I didn’t measure the duty cycle, however.
With my diode and resistor in place my controller works again, I used it 5 times tonight. I also tried to open-circuit the led again in the middle of my tests and noted it did the beeping without action again.
Other bits of interest:
Some UARTs accessible via soldered in headers. These seem to be controlling the separate wireless modules for the tilt sensor and zwave interfaces.
If you jumper the TEST header at power up, the UART to tilt sensor seems to send out a note about being in Factory Test Mode.
RD is relay drive output from micro controller.
SS (switch sense?) seems to be a sense network across the relay contacts, possibly monitoring manual push button presses and/or looking for a minimum voltage on the contacts before operating. My sense network changes to ‘active’ with any more than about 2V across the contacts. Note it is a full wave bridge so polarity insensitive.
You can jumper VPP and COM on the PIC programming header to hold the chip in reset and play with its outputs (like to test the LD signal response or RD without fighting the GPIO output drive)
Many test points are nicely labeled with silkscreen for debugging, such as 12V, 3.3V, Common, etc…
I may order a new CREE led to see if I can get my blinking back, I think there is a compatible part on digikey for 0.75 each (yet, 3.75 shipping).
If anyone wants to sell a unit with this failure mode, I’d consider buying it fairly cheap, I could use a second for my other door.