recommender use promethues sa historyProvider，multicore cpu will only be loaded with one core usage

[SadShark0328]

Which component are you using?:
vpa recommender 0.9.0

vertical-pod-autoscaler

What version of the component are you using?:
0.9.0

Component version:

What k8s version are you using (kubectl version)?:
k8s 1.20

kubectl version Output

$ kubectl version

What environment is this in?:

What did you expect to happen?:

What happened instead?:
In the prometheus and cadvisor versions we use, the reported container_cpu_usage_seconds_total is reported at a single-core granularity

However, recommender loads historical data from prometheus without aggregating these indicators. As a result, only the usage of single-core is loaded and the usage of other cores is discarded

In my case, the physical machine has 8 cores
Therefore, each container pair should have 8 indicators at the same time
· Add debug logs in recommender. It can be seen that 7 indicators are discarded because the time is less than the latest sample time

I0208 14:36:33.018051 1837887 cluster_feeder.go:254] Adding pod {default vpa-admission-plugin-sf-8db91-0} with labels map[] W0208 14:36:33.018083 1837887 container.go:89] sample is valid true, expected: true, sample measure start 2025-02-07 14:36:32.891 +0800 CST. comtainer last CPU sampleStart 2025-02-07 14:36:32.891 +0800 CST, value 4 W0208 14:36:33.018246 1837887 cluster_feeder.go:274] Error adding metric sample for container {{default vpa-admission-plugin-sf-8db91-0} vpa-admission-plugin-phnc-lihulitest}: sample discarded (invalid or out of order) W0208 14:36:33.018256 1837887 container.go:89] sample is valid true, expected: true, sample measure start 2025-02-07 14:36:32.891 +0800 CST. comtainer last CPU sampleStart 2025-02-07 14:36:32.891 +0800 CST, value 4 W0208 14:36:33.018261 1837887 cluster_feeder.go:274] Error adding metric sample for container {{default vpa-admission-plugin-sf-8db91-0} vpa-admission-plugin-phnc-lihulitest}: sample discarded (invalid or out of order) W0208 14:36:33.018265 1837887 container.go:89] sample is valid true, expected: true, sample measure start 2025-02-07 14:36:32.891 +0800 CST. comtainer last CPU sampleStart 2025-02-07 14:36:32.891 +0800 CST, value 4 W0208 14:36:33.018269 1837887 cluster_feeder.go:274] Error adding metric sample for container {{default vpa-admission-plugin-sf-8db91-0} vpa-admission-plugin-phnc-lihulitest}: sample discarded (invalid or out of order) W0208 14:36:33.018293 1837887 container.go:89] sample is valid true, expected: true, sample measure start 2025-02-07 14:36:32.891 +0800 CST. comtainer last CPU sampleStart 2025-02-07 14:36:32.891 +0800 CST, value 5 W0208 14:36:33.018300 1837887 cluster_feeder.go:274] Error adding metric sample for container {{default vpa-admission-plugin-sf-8db91-0} vpa-admission-plugin-phnc-lihulitest}: sample discarded (invalid or out of order) W0208 14:36:33.018304 1837887 container.go:89] sample is valid true, expected: true, sample measure start 2025-02-07 14:36:32.891 +0800 CST. comtainer last CPU sampleStart 2025-02-07 14:36:32.891 +0800 CST, value 5 W0208 14:36:33.018310 1837887 cluster_feeder.go:274] Error adding metric sample for container {{default vpa-admission-plugin-sf-8db91-0} vpa-admission-plugin-phnc-lihulitest}: sample discarded (invalid or out of order) W0208 14:36:33.018315 1837887 container.go:89] sample is valid true, expected: true, sample measure start 2025-02-07 14:36:32.891 +0800 CST. comtainer last CPU sampleStart 2025-02-07 14:36:32.891 +0800 CST, value 5 W0208 14:36:33.018321 1837887 cluster_feeder.go:274] Error adding metric sample for container {{default vpa-admission-plugin-sf-8db91-0} vpa-admission-plugin-phnc-lihulitest}: sample discarded (invalid or out of order) W0208 14:36:33.018326 1837887 container.go:89] sample is valid true, expected: true, sample measure start 2025-02-07 14:36:32.891 +0800 CST. comtainer last CPU sampleStart 2025-02-07 14:36:32.891 +0800 CST, value 5 W0208 14:36:33.018331 1837887 cluster_feeder.go:274] Error adding metric sample for container {{default vpa-admission-plugin-sf-8db91-0} vpa-admission-plugin-phnc-lihulitest}: sample discarded (invalid or out of order) W0208 14:36:33.018341 1837887 container.go:89] sample is valid true, expected: true, sample measure start 2025-02-07 15:36:32.891 +0800 CST. comtainer last CPU sampleStart 2025-02-07 15:36:32.891 +0800 CST, value 5 W0208 14:36:33.018345 1837887 cluster_feeder.go:274] Error adding metric sample for container {{default vpa-admission-plugin-sf-8db91-0} vpa-admission-plugin-phnc-lihulitest}: sample discarded (invalid or out of order) W0208 14:36:33.018348 1837887 container.go:89] sample is valid true, expected: true, sample measure start 2025-02-07 15:36:32.891 +0800 CST. comtainer last CPU sampleStart 2025-02-07 15:36:32.891 +0800 CST, value 4 W0208 14:36:33.018352 1837887 cluster_feeder.go:274] Error adding metric sample for container {{default vpa-admission-plugin-sf-8db91-0} vpa-admission-plugin-phnc-lihulitest}: sample discarded (invalid or out of order)

How to reproduce it (as minimally and precisely as possible):

Anything else we need to know?:

[adrianmoisey]

/area vertical-pod-autoscaler

What version of the component are you using?: 0.9.0
What k8s version are you using (kubectl version)?: k8s 1.20

Both the VPA and Kubernetes versions that you are using are out of support. Is this bug reproducible on the latest of each?