nomad auto scaling using prometheus metric

Question

I'm using Nomad and want to configure a scaling policy in my Nomad file. My current configuration is:

check "allocations-counts" {
    source = "prometheus"
    query  = "nomad_client_allocations_running{node_class=\"myclass\"}"

    strategy "threshold" {
        upper_bound = 3
        delta       = -1
    }
}

After deploying this, it fails with the error: "failed to query source: query returned 4 metric streams, only 1 is expected."

When I run the Prometheus query separately, it returns multiple results like:

• node_class=abcd count=5
• node_class=ebhg count=2
• node_class=kklj count=19
• node_class=erty count=7

What I want to achieve is:

• If any allocation has exactly 2 (e.g., node_class=ebhg with count=2), my AWS ASG should reduce the number of EC2 instances (i.e., delta = -1).

Can anyone suggest the correct configuration for this?