I am trying to get namespace cpu usages usign the following query

"sum_over_time(namespace:container_cpu_usage:sum{{namespace='{namespace}'}}[1d])"

Now if I use the following query and aggregate based on the namespace it is giving me different output.

"sum (sum_over_time(pod:container_cpu_usage:sum{{namespace='{namespace}'}}[1d])) by (namespace)"

There is around 10-15 cores difference between cores.

Namespace_cpu_usage < pod_cpu_usage_aggregated_based_on_namespace (This is what i an getting)

Can anyone explain why is this difference.

I am getting different result for both the following Prometeus Queries, why is the difference there .

"sum_over_time(namespace:container_cpu_usage:sum{{namespace='{namespace}'}}[1d])"

"sum (sum_over_time(pod:container_cpu_usage:sum{{namespace='{namespace}'}}[1d])) by (namespace)"