November 2017
Beginner to intermediate
290 pages
7h 34m
English
Let's start with how Apex determines when state needs to be saved. Instrumental to the process of consistent state saving is the streaming window, which was briefly introduced in Chapter 1, Introduction to Apex. Every streaming window (processing time interval) has a timestamp ID (assigned at the source) and traverses the entire DAG, embedded in the data flow. The engine is aware of these windows and has an opportunity to ask each operator to perform state saving at the same consistent boundary:

The preceding diagram shows how the window markers flow with the data and how checkpoints occur between streaming windows. Normally, ...
Read now
Unlock full access