The job fails silently without any visible errors in the logs, just by reaching some timeout value. Also, adjusting ETL Job Timeout setting does not seem to have any effect.
Release : 15.9.x, 16.x
The cause for this issue seems to be with the Oracle RAC Scan url being used for DWH connectivity. Intermittently it was unable to relay the connection to the actual DB host and hence the job is timing out.
Possible solutions :-