- when scheduling the import don't perform schedule import rounds in a while loop, but do the following:
- schedule one full import round
- schedule a second round where all scrollbased import mediators are triggered
- the decision resolver also fetches decision definitions where the xml hasn't been imported yet. So that he can immediately get in sync.
- decision writer does immediately a refresh on the index after the definitions have been persisted, such that the decision resolver can fetch all definition after one import round
- the IT execution takes less than 10 minutes
The integration tests started to consume a lot of time with up to 20 minutes on jenkins. As a result developers got a bad feedback loop, because they have to wait a long time until they know that the code is working as expected.
We don't need to simulate the original import behavior in the integration tests any longer, since the performance tests cover that this is working as expected and Optimize is importing all engine entities.