Chen Wang

Support in-place resource update feature for VPA

6

**Which component are you using?**: Vertical Pod Autoscaler **Is your feature request designed to solve a problem? If so, describe the problem this feature should solve.**: In Kubernetes 1.25, the...

Exporting Workload Performance Metric as an independent metric that can be queried via Prometheus

2

**Is your feature request related to a problem? Please describe.** Currently, the `pod_energy_stat` metric is a summary object that includes all types of energy consumption metrics, resource usage metrics, and...

Prometheus metrics may have done aggregation.

1

The current load-watcher Prometheus pkg was using the metric of `instance:node_cpu:ratio` to calculate the node utilization However, when this value is still below 60%, I found another metric `instance:node_cpu_utilisation:rate1m` was...

By testing on two bare metal machines, we noticed that CPU frequencies are automatically tuned down when the load is light, regardless of the manual configuration. For example, if we...

The sample error log can be shown below. ``` I0922 05:42:11.413590 1 updater.go:215] evicting pod random-7f8468d595-fvxzf E0922 05:42:11.414606 1 pods_eviction_restriction.go:141] failed to evict pod default/random-7f8468d595-fvxzf, error: Eviction in version "v1"...

Benchmarking script does not limit the maximum concurrency

4

The current benchmarking script if specified with `INF` arrivals, will not limit the maximum concurrency level as shown [here](https://github.com/vllm-project/vllm/blob/703e42ee4b3efed3c71e7ae7d15f0f96e05722d4/benchmarks/benchmark_serving.py#L191). If we can change it to below, we can limit the...