Currently, different SSoTs implement recoverable error handling in different ways. I propose we align on a documented and supported way.
Use Case
If say a system of record has a badly formatted device name that we can't import, we need some way to track this. We don't want to pollute the job logs, because this can be a lot of items depending on the scenario.
Proposed Implementation
Provide some method on an adapter, like add_failed_item or something, that appends to a list. Then, at the end of any job, offer a CSV (example) download of all of these errors.
Environment
Proposed Functionality
Currently, different SSoTs implement recoverable error handling in different ways. I propose we align on a documented and supported way.
Use Case
If say a system of record has a badly formatted device name that we can't import, we need some way to track this. We don't want to pollute the job logs, because this can be a lot of items depending on the scenario.
Proposed Implementation
Provide some method on an adapter, like
add_failed_item
or something, that appends to a list. Then, at the end of any job, offer a CSV (example) download of all of these errors.