Stream Analytics must compute 'average temperature per 5-minute window' over a continuous event stream. Which concept enables bounded calculations on unbounded data?
A.Windowing functions (tumbling/hopping/session)
Time-bounded aggregation = windowing. Correct.
B.Partition keys
Partitioning distributes load — doesn't bound time.
C.Index tuning
Indexing is storage optimization — not streaming logic.
D.Batch checkpoints
Checkpoints are fault recovery — not aggregation.
Why this answer
Streaming queries are infinite — windowing bounds them: tumbling (fixed non-overlapping), hopping (overlapping), sliding, session windows. 'Average per 5 minutes' = tumbling window — the core stream-processing concept.
Correct answer: A · Last reviewed 2026-11-05
Official AWS references
Practice this interactively
Answer without seeing the solution first in the Analytics on Azure practice session.
Practice Analytics on Azure