Update UPLOAD_LOG, CHECK_PROCESSING_TIME, and HANDLE_BATCH_EVENT for multi-task job architecture

[jtherrmann]

See the description for https://github.com/ASFHyP3/hyp3/pull/1166

[jtherrmann]

Steps to verify https://github.com/ASFHyP3/hyp3/pull/1172:

• [x] Confirm the API schema now displays processing_times correctly.
• [x] Confirm processing times list is returned with reasonable values for test jobs.
• [x] Confirm non-zero processing times calculated for a failed job with at least one Batch job attempt
• [x] Confirm processing time is 0 if a Batch job had no attempts.
• [x] Confirm processing times is an empty list if there's an error before the first Batch processing step (i.e. during job type inspection).
• [x] Confirm that the processing_results dict looks correct within the step function JSON.
• [x] Confirm that the PROCESSING_FAILED step works with CHECK_STATUS.
• [x] Confirm that the log for the last Batch job is uploaded. (Edit: I only confirmed this for a job that failed on the first Batch job.)

It would have been nice to verify the expected behavior if there's an error during CHECK_PROCESSING_TIME (processing times remains as an empty list and the error message is recorded to a separate field in the step function JSON), but I couldn't think of an easy way to force that behavior.

[jtherrmann]

TODO:

• [x] Convert processing_time_in_seconds to processing_times in jobs table for all deployments.