Best practice for working with multiple environments and tenants

[KevinDW-Fluxys]

Hi,

We have set up Mimir recently and are wondering what the best practices are for separating data in Mimir tenants.

We have 3 environments (DEV, QUA and PRD) and multiple kubernetes clusters who can belong to one of those environments.
Each cluster also has 2 prometheus instances running, writing data to their own Mimir tenant.
So in Grafana we have a datasource for each prometheus. To clarify this setup, it looks something like this:

PrometheusA-Cluster1-DEV -> MimirTenant-A-1-DEV
PrometheusB-Cluster1-DEV -> MimirTenant-B-1-DEV
PrometheusA-Cluster2-DEV -> MimirTenant-A-2-DEV
PrometheusB-Cluster2-DEV -> MimirTenant-B-2-DEV
PrometheusA-Cluster1-QUA -> MimirTenant-A-1-QUA
PrometheusB-Cluster1-QUA -> MimirTenant-B-1-QUA
PrometheusA-Cluster2-QUA -> MimirTenant-A-2-QUA
PrometheusB-Cluster2-QUA -> MimirTenant-B-2-QUA
...

Would it be better to skip using mimir tenants and send everything to 1 mimir tenant?
How would we filter our metrics between clusters?
Are there any other things to consider?

Thanks for pitching in!