Add retry logic to the Lambdas that make calls to external systems

[briri]

Use Dryad's pattern as a template

https://github.com/CDL-Dryad/dryad-app/blob/main/lib/stash/zenodo_replicate/zenodo_connection.rb

Refer to each API docs to see if they return any rate limit headers on specific errors and use those to determine sleep times.

We will need to incorporate a retry limit though so that the Lambda does not time out