knaufk / enrichments-with-flink Goto Github PK

First of all thank you for your webinar, it was really useful!

Please correct me if I'm wrong, but I think the purpose of EnrichmentFunctionWithPeriodicPreloading is to do one call to loadReferenceData() every 1 hour per thread, while now it's one query every 1 hour per thread X key, because processingTimeTimers are per key.
So if you have lots of keys, say lots of different devices sending iot data, you will hit the db many times.

Maybe a timer is useless in this case and we could simply keep track of the last timestamp when loadReferenceData was called and if X time has passed call it again to refresh data (before processing the current element).

Am I missing something?