Scaling elasticubes? Updating multiple elasticubes? A single reference elasticube?

[mckenziedubois]

We are in the process of converting a single live model into a handful of elasticubes.

Previously, we were using parameters and data groups to point the live model to a database depending on the user that was viewing the sisense report. Now we will need to build a elasticube for each of the databases. We will likely have around 100 elasticubes for all of our databases.

All the elasticubes are identical except for the database they are being pointed to. Schemas, tables, etc, are all the same. So when a change needs to be made, either adding a relationship or perhaps a new table it will need to be rolled out to all elasticubes.

Is there a way we can roll out this change to all the elasticubes? Is there a way to have a base elasticube build that can be referenced and then rolled out to various databases. 

[DRay]

Hi @mckenziedubois,

Thank you for reaching out. I think the Perspectives feature may helpfull. Can you take a look at this documentation and let me know if it helps? https://docs.sisense.com/main/SisenseLinux/perspectives-using.htm

[mckenziedubois]

Thanks for the quick response David, it seems to me that perspectives is more about defining a subset of tables from within the same database. Doesn't seem like I can have individual perspectives point to different dbs but maybe I've overlooked something. 

Ideally, we could have Elasticube_Base, Elasticube A (points to db A), Elasticube B (points to db B)... And then an update to Elasticube_Base could be applied to all Elasticubes. 

[HamzaJ]

Hey @mckenziedubois ,

I do not have real experience with your use case. I do think there is a solution that can make this possible, if you are not afraid of getting your hands dirty with programming. Sisense supports creating/updating datamodels via API (https://sisense.dev/guides/restApi/datamodels.v2.html). 

You could create a template datamodel which has some dynamic variables (like connector configuration), which you can then fill in on the fly while posting it to Sisense. It might take some time in setting up, but I think in the end you could easily push new versions to all the cubes.

Hamza

[mckenziedubois]

Hi Hamza,

During my research today I think I came to the same conclusion that the API is the way to go.  Doesn't look like Sisense has any out of the box solutions. 

Thanks,

[DRay]

Connection Management may also work for you. Here is the doc on that. https://docs.sisense.com/main/SisenseLinux/about-connecting-sisense-with-data-sources.htm

There is also a Sisense Academy course on it. https://academy.sisense.com/connection-management