If you’re putting it into an RRD or something like that, then use a DERIVE or COUNTER data type.  That’ll do it all for you.

 

If it’s the figures I think it is, then it’s basically the number of seconds (or fractions thereof) spent in each state since the system started up.

 

I’ve just used RRD counters to get these graphs… no special manipulation of figures in the data gathering… the percentages are just calculated in the rrdgraph command...rrdtool takes care of working out the difference between the two figures based on the interval since the last set of data.

 

 

From: malar vizhi [mailto:malarsow <at> gmail.com]
Sent: Wednesday, 1 July 2009 4:19 PM
To: net-snmp-users <at> lists.sourceforge.net
Subject: How do I get percentage CPU utilization using ssCpuRawIdle?

 

Hello folks,

Am bit confused to calculate the percentage of ssCpuRawIdle or ssCpuRawNice or ssCpuRawSystem or ssCpuRawUser of host.

I think ssCpuIdle will give exact percentage value which top command shows..but when we see the mib, it is deprecated one. So, We should start of using "raw counters". I came to know from net-snmp forums, need to take two readings of this object, and find the difference between them. That difference divided by the total number of 'ticks' between the two readings (probably 0.01 seconds) will give you the percentage utilization.

And this is the forum where our Dave has replied as above...

http://osdir.com/ml/network.net-snmp.user/2004-02/msg00369.html

Where I have freezed is to calculate the total number of ticks...How should I calculate the total no of ticks?..or will it be always  0.01 seconds for all OS?...

Thanks in Advance....

Malar

Well it’s a basic change over time thing.

 

y – x / t

 

The problem with doing it yourself is you need to store the last value and the time when it was read, and then do the math… if you’re planning to graph that data, then you have to go a step further, and store all the figures

 

 

From: malar vizhi [mailto:malarsow <at> gmail.com]
Sent: Wednesday, 1 July 2009 5:23 PM
To: Lee Standen; net-snmp-users <at> lists.sourceforge.net
Subject: Re: How do I get percentage CPU utilization using ssCpuRawIdle?

 

Thanks Lee....But am not using any third party tool to plot graph....We are using adobe flex for report module(plotting graph). So, those manipulations I have to do manually.It will be better for me If I get any idea on calculating those ticks & all...

Thanks
Malar

On Wed, Jul 1, 2009 at 2:10 PM, Lee Standen <standen <at> staff.iinet.net.au> wrote:

If you’re putting it into an RRD or something like that, then use a DERIVE or COUNTER data type.  That’ll do it all for you.

 

If it’s the figures I think it is, then it’s basically the number of seconds (or fractions thereof) spent in each state since the system started up.

 

I’ve just used RRD counters to get these graphs… no special manipulation of figures in the data gathering… the percentages are just calculated in the rrdgraph command...rrdtool takes care of working out the difference between the two figures based on the interval since the last set of data.

 

 

From: malar vizhi [mailto:malarsow <at> gmail.com]
Sent: Wednesday, 1 July 2009 4:19 PM
To: net-snmp-users <at> lists.sourceforge.net
Subject: How do I get percentage CPU utilization using ssCpuRawIdle?

 

Hello folks,

Am bit confused to calculate the percentage of ssCpuRawIdle or ssCpuRawNice or ssCpuRawSystem or ssCpuRawUser of host.

I think ssCpuIdle will give exact percentage value which top command shows..but when we see the mib, it is deprecated one. So, We should start of using "raw counters". I came to know from net-snmp forums, need to take two readings of this object, and find the difference between them. That difference divided by the total number of 'ticks' between the two readings (probably 0.01 seconds) will give you the percentage utilization.

And this is the forum where our Dave has replied as above...

http://osdir.com/ml/network.net-snmp.user/2004-02/msg00369.html

Where I have freezed is to calculate the total number of ticks...How should I calculate the total no of ticks?..or will it be always  0.01 seconds for all OS?...

Thanks in Advance....

Malar

 

2009/6/8 Ranganathan, Viji (NSN - US/Boca Raton) <viji.ranganathan <at> nsn.com>:

>                      I observe that during snmpwalk, after the
> handler routine with MODE_GET fetches the correct data, the logic goes
> back again to get_first_data_point.

Remember that there is no such thing as an "snmpwalk" request.
This is simply a sequence of GETNEXT requests - one after the other.

For *each* GETNEXT request, the agent will call your 'get_first'
routine (once), followed by your 'get_next' routine (once for each row)
followed by the module handler.

> There are two rows and 24 columns.

> I just want to verify if this many number of calls to the first/next
> routines is normal.

It feels vaguely plausible, yes.

The calculation of exactly how many times each routine should be
called is left as an exercise for the student....

Dave

------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensing option that enables unlimited
royalty-free distribution of the report engine for externally facing
server and web deployment.
http://p.sf.net/sfu/businessobjects
_______________________________________________
Net-snmp-users mailing list
Net-snmp-users <at> lists.sourceforge.net
Please see the following page to unsubscribe or change other options:
https://lists.sourceforge.net/lists/listinfo/net-snmp-users