Methods to spread timeseries across multiple workers for big tenant?

[bbkr (bbkr)]

Hi

I develop application that has huge tenants size differences. And they generate timeseries data ranging from few rows up to 10G rows total. And Citus caught my eye, because it simplifies a lot of things I do manually on current MySQL custom sharding setup.

So I'm familiar with creating tenant based app: https://docs.citusdata.com/en/v11.1/use_cases/multi_tenant.html

I've read also about tenant isolation: https://docs.citusdata.com/en/v11.1/admin_guide/cluster_management.html#tenant-isolation

But what if tenant grows beyond single machine or tenant is unhappy with query performance? Is single machine the end of the road for scaling in tenant based schema? Or maybe there are some methods to distribute timeseries data of single tenant across multiple workers?