I have question regarding failover of a component like QueryDatabaseTable or QueryDatabaseTableRecord.
Suppose, QueryDatabaseTable is configured for some incremental fetch, and the primary node, where QueryDatabaseTable is running, crashes.
The new primary will be selected. Does the new component instance continue to fetch the next increment automatically, meaning, there is a cluster-wide state (max_value_column) of the component instance store somewhere?
Where the value is stored? Internal zookeeper? Can this value extracted and manipulated (e.g. by operations)?