Capture slow query in Cassandra

Author: Sam

Originally Sourced from:

I'm working on some timeout issue in Cassandra production environment and I have set slow_query_log_timeout_in_ms to 5000 ms in Cassandra.yaml to capture any slow queries in debug.log

The following query captured several times as slow query, interestingly there isn't any similar query in the application, so I'm wondering maybe Cassandra translate syntax of slow query to some specific format!

SELECT * FROM WHERE id = b7450619-d2b2-8ba6-0c48-a9bffeef79c0 AND token(user_id, id) > -341911299102519303 AND token(user_id, id) <= -326035898540169815 LIMIT 100>, time 6802 msec - slow timeout 500 msec/cross-node

So my question is how can I capture the actual text of slow query in Cassandra?