The current logic is always resetting the pocs._obs_run_retries count, which means the state machine will go into an endless loop if there are no valid observations but no errors otherwise.
Related to #416, which would fix the issue in this particular case, but the more generic problem could still exist.
The current logic is always resetting the
pocs._obs_run_retriescount, which means the state machine will go into an endless loop if there are no valid observations but no errors otherwise.Related to #416, which would fix the issue in this particular case, but the more generic problem could still exist.