[Cassandra] Visualizations with multiple breakdowns could not be migrated to lens

[rajvi-patel-22]

Description

The Cassandra integration contains two dashboards, [Logs Cassandra] System Logs and [Metrics Cassandra] Overview. While migrating visualizations to lens, most of the visualizations from [Metrics Cassandra] Overview dashboard could not be migrated to the lens. These visualizations contains multiple split series buckets which is not supported in lens yet.

List of panels that contains multiple split series buckets:

1. Disk Usage [Metrics Cassandra]
2. Cache [Metrics Cassandra]
3. Heap Memory [Metrics Cassandra] 
4. Read Latency [Metrics Cassandra] 
5. Write Latency [Metrics Cassandra]  
6. Range Slice Latency [Metrics Cassandra] 
7. Tasks [Metrics Cassandra]  
8. CAS Read/Write Latency [Metrics Cassandra]  
9. Read/Write Failures [Metrics Cassandra]  
10. Dropped Message [Metrics Cassandra]  
11. Read/Write Count [Metrics Cassandra]  
12. Storage Exception [Metrics Cassandra]  
13. Unavailable Requests [Metrics Cassandra]  
14. Request Timeouts [Metrics Cassandra]  

Related Issue

• [Lens] Support for multiple breakdown by kibana#82208

[agithomas]

@rajvi-elastic , going with the comment doesn't the problem get solved with the approach mentioned as per the comment?

[rajvi-patel-22]

@rajvi-elastic , going with the comment doesn't the problem get solved with the approach mentioned as per the comment?

@agithomas According to the comment, multiple breakdowns supports in lens but the behavior is different from the breakdown functionality in visualizations.

The below screenshot compares the behavior of breakdown functionality in both lens and visualizations.

Suppose, the series is grouped by two fields, top 5 values of field1 and top 5 values of field2.

In visualizations, it will first identify top 5 values of field1 and then identifies top 5 values of field2 for those top 5 values of field1.

On the other hand, in lens, first it will group field 1 and field 2 then shows top 5 values from those entire group (field1 + field2).

[agithomas]

In visualizations, it will first identify top 5 values of field1 and then identifies top 5 values of field2 for those top 5 values of field1.

On the other hand, in lens, first it will group field 1 and field 2 then shows top 5 values from those entire group (field1 + field2).

To me, the results / behaviour offered by Lens makes more sense. @SubhrataK , what do you think as the right approach ?

cc @lalit-satapathy , @rameshelastic

[SubhrataK]

@rajvi-elastic Looks like the approach taken by Lens is correct and hence, we should proceed with making the changes.

[rajvi-patel-22]

PR is approved and merged, hence closing this issue:
#5714
#5715