I’m seeing inconsistent behavior between my Scale settings and the Metrics UI for my service, and I’d like clarification on whether this is expected or a bug.

Service / environment (from the URL):

Project ID: fd8e5228-7139-4f3e-ab22-f679fcfeafe0

Service ID: 37e9afd1-2934-44a5-8985-7c3b6ad22df1

Environment ID: e6c594bb-ecf9-480c-b23e-83f029cb234d

What I configured:

I intend to run 2 replicas total, split across two regions (1 replica in US West and 1 replica in Asia Southeast), not 4 replicas.

What I observe in Metrics:

On the Memory chart (Replicas view, last ~6 hours), the legend only lists two series (asia-southeast1-replica-0 and us-west2-replica-0), but the chart clearly shows four distinct horizontal memory lines (two shades per region). The CPU chart also looks like multiple overlapping series.

Billing concern:

Based on my usage breakdown, it appears that Memory usage/cost corresponds to four replicas’ worth of capacity, even though I only configured two replicas as above. Could you please confirm whether billing is correct, and if the four lines on the Memory chart indicate four billable instances, a metrics/UI issue, or something else (e.g. overlapping deploys, multiple processes per replica)?