...
Data Hub requires access to Cloudera Impala, which provides high-performance, low-latency SQL queries on data stored in an Hadoop filesystem (HDFS).
The Data Hub Forwarding Agent bulk loads data in CSV files to HDFS and then inserts it into a Parquet table in the Impala database specified by a Data Hub Profile Profile. The table data is then available for query via Data Hub Query.
In a production environment, it is recommended that the size of the collected files ranges between 1 to 100 MB. Though it is possible to collect and process small batches the overhead of handling a large number of files will have significant impact on performance.
...