Multi-tenant overhead or limit for many tenants

[Tim]

We're considering a setup with many tenants. 

What overhead (RAM/CPU) is there for each additional tenant?

Have others tried working with many tenants (20+)? Did you run into issues?

(I want to know because, depending how we do it, we might have lots of tenants for different environments, products, or other purposes. But only a few would be in use at a time. So I want to know whether we need to resource the server just for the number of users and the build-frequency of datamodels like for a single-tenant server, or if we need to allocate extra RAM/CPU for it to run overheads for all the extra tenants.)

[DRay · Accepted solution]

Hello @Tim,

Thank you for your question. 

The number of tenants does not have a significant impact on the system overhead. The standard best practices still apply for provisioning for the number of concurrent users and build needs.

I hope that helps. Please let us know if you have any other questions, or reach out to your Sisense account team to get connected with a technical resource who can provide more in-depth information.

Have a terrific Tuesday!

David

[Tim]

Hi DRay,

That's what I was hoping to hear, thanks.

Is there a practical limit? Would we run into problems at 100 or 1000 tenants? Even if the server's small because we expect only 5 or so concurrent users and few builds?