The last successful harvesting job (update_database.py) took close to 10 hours, which is mental.
2018-09-04_06:33:08 Starting full monument update. 2018-09-04_06:33:08 Load changes to monuments_config... 2018-09-04_06:33:23 Recreating the source tables... 2018-09-04_06:33:37 Full source database update... 2018-09-04_17:03:21 Update monuments_all table... 2018-09-04_17:58:27 Make statistics...
Each dataset harvest being independent, we should consider parallelizing the processing.