When usage is processed by swatch-producer-aws, it should increment a counter named "swatch_producer_metered_total" with the amount in metric units, having labels:

• product (swatch product tag)
• metric_id (metric ID)
• billing_provider
• status (failed/succeeded/duplicate)

• What about retry: we only care about the final state after AWS properly returns 

Done:

• Metric visible in prometheus/grafana explorer (dashboard will come later)
• QE verify that the pods are capturing these metrics via :9000/metrics or stage prometheus instance.