Which data type should be used to store data received from model-driven telemetry?

BigQuery is an analytical data warehouse optimized for batch queries on large datasets, not for continuous high-frequency telemetry ingestion.

Time series databases such as InfluxDB or Prometheus are optimized for ingesting high-frequency, timestamped data points - exactly what model-driven telemetry produces via streaming subscriptions. They provide efficient compression, retention policies, and time-range queries that relational and general-purpose NoSQL databases cannot match for this workload.

NoSQL databases are general-purpose document or key-value stores and lack the time-indexed optimization required for streaming telemetry data.

PostgreSQL is a relational database not natively optimized for time-series ingestion patterns, high write throughput, or automatic data expiration typical in telemetry use cases.