Stream Processing Semantics: Time, State, and What Exactly-Once Actually Covers
In streaming, a number is never simply right. It is right as of a boundary you chose, over data that had arrived by then, and it may change tomorrow.
-
-
Two-Phase Commit Sink
An engine’s exactly-once guarantee stops at its own state. A sink that commits with the checkpoint is how that guarantee reaches anything outside it.
-
The Job Succeeded and the Numbers Are Wrong: Duplicates, Loss, and Reprocessing
Every job succeeded and the total is still wrong. Here is where duplicates and missing rows come from, what “exactly once” really promises, and how to verify.
-
Delivery Semantics
Three promises about how many times a record arrives. The useful question is not which one a product advertises, but where its guarantee stops.