Tabular query resource limits and intensive use

[brownag]

RE: get_rap_production16day_table()

Assuming the coordinates are valid, the core issue seems to be server-side: under some conditions server does not generate a valid JSON response. This could be due to malformed input (as was the case with empty geoms), or some other internal failure.

There may be rate-limiting or resource limit you are hitting with large number of requests. #14 is a similar case where we run into issues with network intermittency in getting remote files with GDAL. This issue can be difficult to debug as it depends on network, pattern of use, user/project/server load, etc. and is not a "defined" behavior for the API.

If after #16 you are getting many feature IDs throwing warnings/NULL results, try to batch your AOIs into smaller chunks (say, ~1000 points each), with a brief period of waiting between each chunk, and then combine the chunked results. If doing that resolves your issue, then that might suggest there are implicit rate limitations we could better address in the function when user input is large.

Originally posted by @brownag in #12