Something in OWS makes exceptions disappear

[pjonsson]

While debugging #1386, I got a lot of strange one-line/word error message in my log, and I'm pretty sure there is something inside OWS that takes exceptions and makes them disappear, and then logs the message from the exception. I'm not sure what though, but if I had to guess I would blame the code at

datacube-ows/datacube_ows/ogc.py

Line 185 in aeb2a92

except Exception as e: # pylint: disable=broad-except

.

It's possible that code is only missing a _LOG.exception(e) so we could see crashes and then start to catch the unexpected exceptions in a better scope.

Edit: #1388 is an example of what makes me believe something inside OWS is doing something with exceptions, the only two mentions of source code in OWS are these, and those are for the caused exception that starts with

Traceback (most recent call last):
  File "/app/lib/python3.12/site-packages/datacube_ows/ows_configuration.py", line 1178, in force_range_update
    self._ranges = self.ows_index().get_ranges(self)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/app/lib/python3.12/site-packages/datacube_ows/index/api.py", line 236, in inner
    dc = Datacube(env=dcl.local_env)
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^

but ows_configuration.py:1178 isn't an entry function, so there are callers missing from the backtrace.

[pjonsson]

Here's one of the culprits found at the end of ows_configuration.py:

def get_config(
    refresh: bool = False, called_from_update_ranges: bool = False
) -> OWSConfig:
    cfg = OWSConfig(
        refresh=refresh, called_from_update_ranges=called_from_update_ranges
    )
    if not cfg.ready:
        with contextlib.suppress(ODCInitException):
            cfg.make_ready()
    return cfg

@SpacemanPaul do we do what we can to fix the callers and then rip off the band-aid and play whack-a-mole for a while with the crashes we missed, or do you have any other ideas?