Verify API behavior when a subset of the EDS request fails

[brucecrevensten]

The goal here is to study the code path to ensure that any HTTP errors or data failures from Rasdaman bubble up to trigger an error status on the cumulative /eds endpoint. This will ensure that cache repopulation can succeed.

[brucecrevensten]

This is definitely real and keeps happening but we don't have any kind of exact knowledge of what is happening. Probably Temperature is a challenge here.