Prove fleet scale-out with controller state, not summed process gauges

During rolling deployments, old and new processes for the same logical worker can export metrics at the same time. Summing per-process active calls or capacity can therefore create a false fleet peak and a false impression that autoscaling added workers. Compare unique worker identities and deployment generations, then confirm desired, actual, and routeable counts from the controller and hosting platform. Treat overlap windows separately from ordinary steady-state samples. A higher configured maximum is only a ceiling; it is not evidence that scale-out succeeded.

Applies when: rolling replacements overlap in monitored autoscaled fleets. Keywords: autoscaler duplicate metrics rolling deployment active calls scale out proof