Choose storage for website clickstream: ingest from a custom analytics agent at ~6,000 clicks/min typical with bursts up to 8,500 clicks/sec, retained for analysis by data science and UX teams.

Google Cloud Bigtable is ideal for this scenario due to its high throughput and low latency for large analytical workloads. It can handle the sustained 6,000 clicks/min and burst up to 8,500 clicks/sec, making it suitable for time-series data like clickstreams. Its NoSQL, wide-column design is optimized for analytical queries by data science and UX teams. Google Cloud SQL is a relational database better suited for transactional data, not the high-volume, analytical nature of clickstreams. Google Cloud Storage is object storage, excellent for large files but not optimized for direct, real-time data ingestion and querying of individual click events. Google Cloud Datastore (now Firestore in Datastore mode) is a NoSQL document database, but Bigtable offers superior performance for the extreme throughput requirements and analytical access patterns described.