telemetry-tempest-plugin prometheus autoscaling scenario should use the metering.server_group for instance grouping instead of instance names. This will likely be just a simple change to the format of the query.

We might also want to do a telemetry-tempest-plugin release after this to get these changes downstream.