Standardize log data for services running on GKE using Google-recommended practices and minimize steps. Which actions should you take? (Choose two.)

Writing log output to stdout as single-line JSON is the recommended and most straightforward way to get structured logs into Cloud Logging from GKE. The Cloud Logging agent automatically collects stdout and stderr streams from containers and ingests them. Using single-line JSON ensures the logs are parsed as structured entries, making them queryable and analyzable. Creating aggregated exports on application logs to BigQuery facilitates log analytics by providing a powerful, scalable data warehouse for querying and analyzing large volumes of log data. This is a standard practice for long-term retention and advanced analysis of logs, once they are ingested into Cloud Logging. Mandating the use of the Logging API in application code adds unnecessary complexity and couples the application to a specific logging implementation. Cloud Storage is less suitable than BigQuery for direct log analytics due to its object storage nature, which is not optimized for querying structured data. Using Pub/Sub and Dataflow introduces significant overhead for a common logging task that Cloud Logging handles natively.