pcp
[Top] [All Lists]

Re: [pcp] kernel.all.uptime semantics

To: pcp@xxxxxxxxxxx
Subject: Re: [pcp] kernel.all.uptime semantics
From: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu, 19 May 2016 09:55:37 +1000
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <573C4370.7040707@xxxxxxxxxx>
References: <573C4370.7040707@xxxxxxxxxx>
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.7.2
On 18/05/16 20:26, Marko Myllynen wrote:
Hi,

$ pminfo -d kernel.all.uptime

kernel.all.uptime
     Data Type: 32-bit unsigned int  InDom: PM_INDOM_NULL 0xffffffff
     Semantics: instant  Units: sec
$

Is there a reason why this is instant not a counter?

Yep.

Most often, people are interested in how long a system has been up, not the fact that it has been up 100% over the last sample interval (which is what would be reported if you used pmval/pmie/pmchart to report this, for example).

It is otherwise an arbitrary decision.

Other "time" counters, like the CPU time do have the semantics of counters because their rate is usually less than 100% and people are more interested in the recent rate than the total time consumed since boot.

<Prev in Thread] Current Thread [Next in Thread>