Dear @mludwig2,
The problem is actually with USGS and some wrong setting on their side where the data is stored. We reported this problem to them and it usually gets fixed within a matter of weeks.
Looking at your pg I see that you are requesting a 29 time range. Within such a long time range it can always be that one of the hundreds of products available are broken. I would therefore recommend you to split your temporal extend into several small ones and for now exclude the falsely date on 26.02.1986.
In addition you could enable the “soft error” feature in python so that a single failing file doesn’t ruin the whole job (see this related topic for more more details).
Best,
Daniel