The Performance service in Opscenter shows slow queries by selecting from the table dse_perf.node_slow_log. However not all records in the node_slow_log table are listed in the Opscenter UI.
Opscenter offers the possibility to log slow queries by enabling the slow queries facility. This feature allows the user to quickly get an indication about the queries that violate a configurable threshold: see threshold_ms in dse.yaml.
However, Opscenter will not show the entire list of slow queries as they are recorded in the dse_perf.node_slow_log table but only a small subset.
The table dse_perf.node_slow_log could potentially contain 1000s of records and most of the records could also be duplicates. Opscenter is designed to only scan the table for changes that have occurred in a predefined time interval (not configurable by the user), and only shows these records in Service/Performance.
This is how Opscenter is designed to work to limit the impact on performance by the Performance service itself. A request for improvement has been filed to add a message in Opscenter UI or Doc to make it clearer to users about what they see in Performance Service.
Internal request for improvement: OPSC-8574