Handheld scanners began sending PII into Kafka topics; you must quickly stop PII reaching analytics using scalable, managed cloud services. Which solution do you implement?

The correct solution leverages Cloud Functions and Cloud DLP for a scalable, managed, and immediate response to PII in Kafka. A Cloud Function can be triggered by new messages in Kafka topics, allowing real-time processing. Cloud DLP is specifically designed for identifying, classifying, and redacting sensitive data, offering robust tagging and confidence levels. This allows for immediate action: either passing clean data or quarantining PII for review, preventing it from reaching downstream analytics. Creating an authorized view in BigQuery is a good security measure for data at rest but doesn't prevent PII from entering the pipeline. Installing a third-party tool on Compute Engine adds operational overhead and isn't as natively scalable or managed as Cloud DLP. Cloud Logging analyzes data after it has passed through, which is too late to prevent PII from reaching analytics.