When defining a logger table, according to the documentation, only a limited number of rows is loaded in the SLProtocol process.
Does this mean that only that limited amount of rows is also stored in the elementdata table?
Or is there never data stored in the elementdata table when using a logger table?
In case there's partial data stored in the elementdata table when using a logger table, should we combine this with the volatile option?
When using a logger table, in what database table can we find the table's data?
Logger tables are stored in a separate table in the DB created specifically for that element. So it should not be stored in the elementdata table in DB as it is already in a custom table in DB. When opening the element in Cube you will want to see the content of that table asap, this is why a part of the data is stored in the SLProtocol process, so that part can already be returned to the client while other parts can be loaded later through lazy loading if needed. So only a part to avoid blowing up memory in case of large logger tables. That part is in memory to still have reasonable responsiveness when retrieving the table through lazy loading.
The logger table can be stored in Cassandra and/or Elastic. Depending on your use case you might want to choose one over the other. I would recommend using Elastic if there might/can be many elements on a DMS using that logger table. As Cassandra can't cope well with many DB tables (>200).