GCP Professional Data Engineer Practice Question

Cloud Bigtable is purpose-built for high-volume, single-keyed workloads such as time-series telemetry. It delivers single-digit millisecond latency for point reads or contiguous key-range scans, can ingest hundreds of thousands of writes per second, and supports cluster autoscaling with no downtime. A composite row-key such as meter_id#timestamp keeps recent data for each meter in the same locality, satisfying the 10 ms SLA. BigQuery is optimized for analytical scans and typically returns interactive rather than sub-10 ms responses. Cloud Spanner offers strong transactional semantics but its cost-per-throughput and read latency are higher for simple key-value access patterns. Firestore cannot sustain the required write throughput or latency at this scale. Therefore, Cloud Bigtable is the most appropriate choice.