Details
-
Task
-
Status: Done
-
L3 - Default
-
Resolution: Fixed
-
None
-
None
Description
AT:
- the counts of historic variable instances/process instances/activity instances are cached for each process definition
- a single process definition instance count in the cache is updated as soon as the process definition based index handler fetches instances for a single process definition
- ScrollBasedIndexHandler do not update the count in every import round but only, when the progress is calculated
- ScrollBasedIndexHandler do not perform a request to the engine on update count, but rather check elasticsearch for that
Hint:
right now we always fetch the total count of variable/activty/process instances. However, this gets very expensive (especially in postgres) with huge amount of data, which results that the import does not continue and the import progress returns wrong results.