diff --git a/docs/proposals/20220228-advanced-cpuset-manger.md b/docs/proposals/20220228-advanced-cpuset-manger.md index c60b576ae..8386112b2 100644 --- a/docs/proposals/20220228-advanced-cpuset-manger.md +++ b/docs/proposals/20220228-advanced-cpuset-manger.md @@ -34,7 +34,7 @@ This proposal provides a new mechanism to manage cpusets, which allows sharing c ## Motivation -Some latency-sensitive applications have lower lantency and cpu usage when running with specific cores, which results in fewer context switchs and higer cache affinity. +Some latency-sensitive applications have lower lantency and cpu usage when running with specific cores, which results in fewer context switches and higer cache affinity. But kubelet will always exclude assigned cores in shared cores, which may waste resources.Offline and other online pods can running on the cores actually. In our experiment, for the most part, it is barely noticeable for performance of service. ### Goals diff --git a/docs/proposals/20220712-recommendation-framework-internal.md b/docs/proposals/20220712-recommendation-framework-internal.md index aa7d40b11..88ad74737 100644 --- a/docs/proposals/20220712-recommendation-framework-internal.md +++ b/docs/proposals/20220712-recommendation-framework-internal.md @@ -18,7 +18,7 @@ In order to solve the above problems, we hope the whole recommendation process i ## Goals - Define the architecture of Recommendation Framework. -- Define the interfaces of Recomendation Framework Internal modules. +- Define the interfaces of Recommendation Framework Internal modules. ## Non-Goals diff --git a/docs/tutorials/using-effective-hpa-to-scaling-with-effectiveness.md b/docs/tutorials/using-effective-hpa-to-scaling-with-effectiveness.md index b0c5770e9..cf29c0830 100644 --- a/docs/tutorials/using-effective-hpa-to-scaling-with-effectiveness.md +++ b/docs/tutorials/using-effective-hpa-to-scaling-with-effectiveness.md @@ -346,9 +346,9 @@ EffectiveHorizontalPodAutoscaler is compatible with HorizontalPodAutoscaler(Whic With EHPA, users can configure CronMetric、PredictionMetric、OriginalMetric at the same time. -**We highly recomend you configure metrics of all dimensions. They are represtenting the cron replicas, prior predicted replicas, posterior observed replicas.** +**We highly recommend you configure metrics of all dimensions. They are represtenting the cron replicas, prior predicted replicas, posterior observed replicas.** -This is a powerful feature. Because HPA always pick the largest replicas calculated by all dimensional metrics to scale. Which will gurantee your workload's QoS, when you configure three types of autoscaling at the same time, the replicas caculated by real metric observed is largest, then it will use the max one. Although the replicas caculated by prediction metric is smaller for some unexpected reason. So you don't be worried about the QoS. +This is a powerful feature. Because HPA always pick the largest replicas calculated by all dimensional metrics to scale. Which will guarantee your workload's QoS, when you configure three types of autoscaling at the same time, the replicas caculated by real metric observed is largest, then it will use the max one. Although the replicas caculated by prediction metric is smaller for some unexpected reason. So you don't be worried about the QoS. #### Mechanism diff --git a/site/content/en/docs/Proposals/20220228-advanced-cpuset-manger.md b/site/content/en/docs/Proposals/20220228-advanced-cpuset-manger.md index 633f87ad5..56aa26d8f 100644 --- a/site/content/en/docs/Proposals/20220228-advanced-cpuset-manger.md +++ b/site/content/en/docs/Proposals/20220228-advanced-cpuset-manger.md @@ -27,7 +27,7 @@ This proposal provides a new mechanism to manage cpusets, which allows sharing c ## Motivation -Some latency-sensitive applications have lower lantency and cpu usage when running with specific cores, which results in fewer context switchs and higer cache affinity. +Some latency-sensitive applications have lower lantency and cpu usage when running with specific cores, which results in fewer context switches and higer cache affinity. But kubelet will always exclude assigned cores in shared cores, which may waste resources.Offline and other online pods can running on the cores actually. In our experiment, for the most part, it is barely noticeable for performance of service. ### Goals diff --git a/site/content/zh/docs/Proposals/20220228-advanced-cpuset-manger.md b/site/content/zh/docs/Proposals/20220228-advanced-cpuset-manger.md index 633f87ad5..56aa26d8f 100644 --- a/site/content/zh/docs/Proposals/20220228-advanced-cpuset-manger.md +++ b/site/content/zh/docs/Proposals/20220228-advanced-cpuset-manger.md @@ -27,7 +27,7 @@ This proposal provides a new mechanism to manage cpusets, which allows sharing c ## Motivation -Some latency-sensitive applications have lower lantency and cpu usage when running with specific cores, which results in fewer context switchs and higer cache affinity. +Some latency-sensitive applications have lower lantency and cpu usage when running with specific cores, which results in fewer context switches and higer cache affinity. But kubelet will always exclude assigned cores in shared cores, which may waste resources.Offline and other online pods can running on the cores actually. In our experiment, for the most part, it is barely noticeable for performance of service. ### Goals