The Streamshub Flink SQL runner pulls in the flink-s3-fs-hadoop library in order to provide access to s3 filesystems in Flink SQL.

The hadoop libaray is vulnerable to several CVEs. There is another library based on presto, flink-s3-fs-presto, which should provide the same functionality but avoid using the vulnerable Hadoop libraries.

We should verify we can have the same functionality, including checkpointing and savepointing, using the presto library and switch these out for the next release of the SQL runner.