reorg cleanup query timeout

[ryandotsmith]

It is possible that the query used to delete data during a reorg is too slow for the default 10s database timeout. During a reorg Shovel will delete data from shovel.task_updates with this query and all the integration's tables with this query.

It is possible that one of these two queries is taking longer than 10s. Since Shovel automatically creates an index on shovel.task_updates, I'm wondering if we are missing an index on the integration tables.

The ideal fix is to make these queries fast. Perhaps with a better index.

If you are experiencing this problem please use the postgres logs to show the exact query that is timing out. Also please share the table DDL so we can look for missing indexes.