Summary
In order to vary the data generated by the default insert operation defined in the cppsuite in the database_operation class, we could randomise which collections the threads are working on.

Currently, each thread is assigned to a defined set of collections and inserts some key/value pairs into each of them, we end up having a uniform distribution of the data across all collections.

Instead, we could use different pseudo-random generations. The chosen distribution could be specified through the configuration file. This cpp documentation lists the different ones available in c++.

Motivation

• Is this issue urgent?
  It's always good the generate random sets of data to test more workloads.

Acceptance Criteria (Definition of Done)

• Testing
  We can specify a configuration through the test configuration file and the insert operation uses it to insert data in the collections. Once this is done, it can be applied to the other operations (if this work ends up being lots of changes, creating new tickets is recommended)
• Documentation update
  Check the documentation written as part of WT-9109 and WT-9111