Depending on the source, we have to use the right GCP service. We can also use Apache Flume/Apache Storm/Apache Kafka/Apache Beam for streaming data - services from Hadoop ecosystem. Interesting things is that Apache Beam is an alternative to Cloud Data flow as well.
To ingest data from Cloud SQL, we can use the Sqoop component of Hadoop. And for batches we have plenty of options - Apache Flume can be one of those as well.