You’re writing user behavior events to BigQuery via the BigQuery Storage Write API and must prevent duplicates using the simplest operational approach. What do you do?

The BigQuery Storage Write API's committed write stream type provides exactly-once delivery semantics, automatically preventing duplicate records from being written to BigQuery. This is the simplest operational approach because it handles deduplication internally within BigQuery, requiring no additional infrastructure or custom logic. Creating a write stream in the default type (pending) requires a separate FlushRows operation and does not inherently prevent duplicates. Configuring a Kafka cluster or a Pub/Sub topic with Cloud Functions introduces significant operational overhead and complexity for deduplication, as these solutions require external mechanisms to track and filter duplicates, making them less operationally simple than BigQuery's built-in committed stream.