Your Browser is Out of Date

Nytro.ai uses technology that works best in other browsers.
For a full experience use one of the browsers below
PowerScale: Beginner’s Guide to PowerScale KPI Performance Metrics
Executive summary Monitoring a cluster Metrics KPI summary
Introduction
References

Introduction

Introduction

  • The following tables provide a KPI summary.

    Table 4.      Protocol latency

    Protocol

    High performance

    Medium performance

    General performance

    SMB

    <= 10 ms

    <= 15 ms

    <= 20 ms

    NFS

    <= 5 ms

    <= 10 ms

    <= 20 ms

    HDFS

     

     

    <= 10 seconds

    Table 5.      Workload balance

    Variance between nodes

    Ideal

    Acceptable

    Average network throughput/node

    +/- 3%

    +/- 10%

    Client count (connected and active)

    +/- 3%

    +/- 10%

    Table 6.      Connection counts per node

    Protocol

    Active clients per node

    Connected clients per node

    NFSv3

    1000

    Limited by available sockets

    NFSv4

    1000

    Limited by available sockets

    SMB1

    1000

    2000

    SMB2

    3000

    27000

    SMB3

    3000

    27000

    Table 7.      General metrics

    General

    Ideal

    Good

    Acceptable

    Disk I/O queue time

    <= 1 ms

    <= 4 ms

    <= 8 ms

    CPU utilization/node

    N/A

    N/A

    <= 60%

    External network errors

    0

    N/A

    < 1/day average

    Cluster storage capacity

    <= 80%

    N/A

    <= 85%