We often procedure join to an API call, essentially looping through and calling an endpoint for each row in another dataset, and caching the resulting dataset. We often have thousands of rows that we need to loop through. Currently if one of those fails due to an API 404 type error or a network connection error, the entire cache job fails. We would like to have the option to retry that call or skip it. We tried wrapping the API call in a custom procedure to be able to add try/except type logic. The current EXCEPTION keyword seems to only hand SQL exceptions. When the API call fails, it seems to be java exception, and the exception keyword isn't able to catch it. This causes us to have to either manually rerun caches when they fail or to add extra caching layers to handle cache jobs that loop thousands of times and fail in the middle because of a network connectivity issue.