This implements *simplistic* reloading of ProcessingTypeData - treat it as experimental/working PoC
One of the biggest issues is that it can break current operations - when reloadAll is invoked, e.g. during
process deploy via FlinkRestManager it may very well happen that http backed is closed between two Flink invocations.
To handle this correctly we probably need sth like:
def withProcessingTypeData(processingType: ProcessingType)(action: ProcessingTypeData=>Future[T]): Future[T]
to be able to wait for all operations to complete
Another thing that needs careful consideration is handling exception during ProcessingTypeData creation/closing - probably during
close we want to catch exception and try to proceed, but during creation it can be a bit tricky...
This implements *simplistic* reloading of ProcessingTypeData - treat it as experimental/working PoC
One of the biggest issues is that it can break current operations - when reloadAll is invoked, e.g. during process deploy via FlinkRestManager it may very well happen that http backed is closed between two Flink invocations. To handle this correctly we probably need sth like: def withProcessingTypeData(processingType: ProcessingType)(action: ProcessingTypeData=>Future[T]): Future[T] to be able to wait for all operations to complete
Another thing that needs careful consideration is handling exception during ProcessingTypeData creation/closing - probably during close we want to catch exception and try to proceed, but during creation it can be a bit tricky...