After a node failure, a parking-citation streaming job must resume without losing its rolling window state. What mechanism supports that recovery?
Select an answer to reveal the explanation.
Short Explanation
Checkpoints are like saving your game before a tough boss. When the streaming job crashes, it reloads from that save instead of pretending the whole shift never happened.
Full Explanation
Streaming engines periodically checkpoint offsets and application state to durable storage. After failure, the job can restore from the latest checkpoint and continue with consistent window or aggregate state. For civic workloads, checkpointing is a core fault-tolerance practice rather than an optional luxury.