Latency issues with cluster and resource pool metrics

Products

• VMware vCenter Server

Issue/Introduction

Symptoms:

• Last 20-30 minutes data of a handful metrics (e.g. CPU performance/usage) for cluster and resource pool are lacking.
  
  
• The affected performance charts are:
  
  
  • Cluster historic overview/advanced charts (cpu usage mhz, consumed mem, granted mem, active mem, balloon mem, swap used).
    
    
  • Resource pool historic overview/advanced charts.
    
    
  • topN charts

Environment

VMware vCenter Server 6.5.x
VMware vCenter Server 6.0.x
VMware vCenter Server 5.5.x
VMware vCenter Server Appliance 6.0.x
VMware vCenter Server Appliance 5.5.x
VMware vCenter Server 6.7.x

Cause

As part of the performance improvements in vSphere 5.5 Update 3, some cluster and resource pool metrics that appear in performance charts are displayed faster in the vSphere Web Client user interface. As a result of this, if your vCenter Server version is 5.5 Update 3 or later, the last 30 minutes data of following performance chart for cluster and resource pool are lacking.

This is a functional impact for the used pre-aggregation which is introduced in vCenter Server 5.5 Update 3.

Resolution

This is a known issue affecting the vCenter Server 5.5 Update 3 or later.

To resolve this issue, add/edit the value of config.vpxd.stats.collectPreAggregatedIntervalFactor to '1', from default value '4' and restart vpxd.

Note: These metrics may be 20-30 minutes behind on a daily basis. By default it runs every 4 daily intervals (default 5 minutes), i.e. every 4 * 5 = 20 minutes. If the setting is set to N(N=1,2,3,.), then the hosts will be queried every (N x 5) minutes and the latency in the pre-aggregated metrics will be smaller.

Additional Information

群集和资源池衡量指标的延迟问题
クラスタおよびリソース プールのメトリックに関する遅延の問題