Forum Discussion

Yehey_100134's avatar
Yehey_100134
Icon for Nimbostratus rankNimbostratus
Oct 28, 2011

LTM Performance Stats

Hi,

 

 

 

I was assigned the uneviable task to gather performce stats for our LTM on a dialy basis using the Enterprise Manager. I would gether the below stats and tabulate them then email it to the specified recipients. But this becoming cumbersome and sometimes fail to accomplish it. Im now planning on automating the process.

 

 

 

CPU (Max)

 

Memory (Max)

 

Active Connection (Max)

 

Throuhput (Min and Max of Server and client)

 

 

 

My questions is what would be easier to implement, rrd or snmp? I'm not familiar with both. I was able to get the rrd file but im having problems extracting/reading the logs. I tried snmp but it seems its difficult to gether historical logs (1 day) stats using snmp.

 

 

 

Really appreciate your comments and suggestions.

 

 

 

 

 

Thanks,

 

 

 

Yehey

 

management
monitoring
  • Kurt_Knochner_5's avatar
    Kurt_Knochner_5
    Icon for Cirrus rankCirrus
    Oct 28, 2011
    My questions is what would be easier to implement, rrd or snmp? I'm not familiar with both. I was able to get the rrd file but im having problems extracting/reading the logs. I tried snmp but it seems its difficult to gether historical logs (1 day) stats using snmp.

     

    rrd is a database to store performance data (or other values). snmp is a protocol to collect data from the LB. So it's not rrd or snmp, it's rather both.

     

     

    Please take a look at cacti. It creates nice graphs, offers a decent web gui and there is a lot of information available at devcentral. Just search for cacti.

     

     

    http://www.cacti.net/

     

     

    With the nectar plugin you can even send reports via e-mail (http://docs.cacti.net/plugin:nectarpurpose).

     

     

    Regards

     

    Kurt Knochner

     

  • Josh_41258's avatar
    Josh_41258
    Icon for Nimbostratus rankNimbostratus
    Nov 07, 2011
    My suggestion is to use a third party product/NMS (ZenOSS, etc) to poll this data via SNMP. Once the NMS has the data, it will likely graph it and allow you to report on it as well.