Currently, conductor treats all errors as one type and then increase the retry count if feasible to retry.
In some use cases, users want to differentiate errors with different types (e.g. user error, platform failure, etc.) and then set different retry policies accordingly.
Currently, conductor treats all errors as one type and then increase the retry count if feasible to retry. In some use cases, users want to differentiate errors with different types (e.g. user error, platform failure, etc.) and then set different retry policies accordingly.