Layer7 API Management

Expand all | Collapse all

Best practice on 100% of "Request Routed"

  • 1.  Best practice on 100% of "Request Routed"

    Posted 06-20-2018 02:59 PM

    I am curious to know that what is the criteria to decide for deciding additional node.

    here are clustering environment dashboard screen shots.

     

    Does latency slowness(performance) is related to 100% utilization of "Request Routed"? 

    what are the criteria to decide multiple nodes?



  • 2.  Re: Best practice on 100% of "Request Routed"

    Posted 06-20-2018 07:16 PM

    Tattwa

     

     

    Looking at your screenshot, the load on the cluster is very low. Load average is determined as high or low based on the number of cpu's on the server. Load average of above 4 on a 4 CPU machine is considered high. 

     

    Refer Gateway Dashboard - CA API Gateway - 9.2 - CA Technologies Documentation  for details on the values displayed on policy manager dashboard.

     

    Load Sharing %
    Indicates the percentage of total cluster traffic being handled by the node over the past 60
    seconds, expressed both as a percentage and as a dynamic bar graph.
    A value of "0" indicates no activity.
    Request Routed %
    Indicates the percentage of current routing activity being handled by the node over the
    past 60 seconds, expressed both as a percentage and as a dynamic bar graph.
    A value of "0" indicates no activity.
    Load Avg

    The average number of work processes completed over the last 60-second period. For
    Appliance Gateway installations, values reaching or exceed the number of CPUs on the
    appliance indicates that the Gateway is under heavy load and overall server performance
    is slow. For example, a value of 4.0 or greater on a 4-CPU appliance indicates heavy load
    on the Gateway.
    This setting does not apply to Software Gateway installations.



  • 3.  Re: Best practice on 100% of "Request Routed"

    Posted 12-17-2018 11:47 AM

    Good afternoon,

     

    Were you able to resolve the issue? What was the final result?

     

    Sincerely,


    Stephen Hughes
    Broadcom Support