<div dir="ltr">This is about accumulative values, I&#39;m also asking about stats like CPU usage of the VM\Host that is not reported in absolute value.<div>Can you bump the thread?</div></div><div class="gmail_extra"><br clear="all"><div><div class="gmail_signature" data-smartmail="gmail_signature"><div dir="ltr"><div><div dir="ltr"><pre cols="72"><span style="font-family:arial,helvetica,sans-serif">Yaniv Dary
Technical Product Manager
Red Hat Israel Ltd.
34 Jerusalem Road
Building A, 4th floor
Ra&#39;anana, Israel 4350109

Tel : +972 (9) 7692306
        8272306
Email: <a href="mailto:ydary@redhat.com" target="_blank">ydary@redhat.com</a>
IRC : ydary</span></pre>
</div></div></div></div></div>
<br><div class="gmail_quote">On Mon, Feb 27, 2017 at 10:11 AM, Francesco Romani <span dir="ltr">&lt;<a href="mailto:fromani@redhat.com" target="_blank">fromani@redhat.com</a>&gt;</span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
  
    
  
  <div bgcolor="#FFFFFF" text="#000000"><div><div class="h5">
    <br>
    <div class="m_-7423557398579924391moz-cite-prefix">On 02/26/2017 03:13 PM, Yaniv Dary
      wrote:<br>
    </div>
    <blockquote type="cite">
      <div dir="ltr"><br>
        <div class="gmail_extra">
          <div class="gmail_quote">
            <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
              <div bgcolor="#FFFFFF" text="#000000">
                <div>
                  <div class="m_-7423557398579924391h5">
                    <blockquote type="cite">
                      <div dir="auto">
                        <div>
                          <div class="gmail_extra">
                            <div class="gmail_quote">
                              <blockquote class="m_-7423557398579924391m_-684850395146887592quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> 2.
                                collectd *intentionally* report metrics
                                as rates, not as absolute<br>
                                values as Vdsm does. This may be one
                                issue in presence of restarts/data<br>
                                loss in the link between collectd and
                                the metrics store.<br>
                              </blockquote>
                            </div>
                          </div>
                        </div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">How does this work? </div>
                        <div dir="auto">If we want to show memory usage
                          over time for example, we need to have the
                          usage, not the rate. </div>
                        <div dir="auto">How would this be reported? <br>
                        </div>
                      </div>
                    </blockquote>
                    <br>
                  </div>
                </div>
                I was imprecise, my fault.<br>
                <br>
                Let me retry:<br>
                collectd intentionally report quite a lot of metrics we
                care about as rates, not as absolute values.<br>
                Memory is actually ok fine.<br>
                <br>
                  a0/virt/disk_octets-hdc -&gt; rate<br>
                  a0/virt/disk_octets-vda<br>
                  a0/virt/disk_ops-hdc -&gt; rate<br>
                  a0/virt/disk_ops-vda<br>
                  a0/virt/disk_time-hdc -&gt; rate<br>
                  a0/virt/disk_time-vda<br>
                  a0/virt/if_dropped-vnet0 -&gt; rate<br>
                  a0/virt/if_errors-vnet0 -&gt; rate<br>
                  a0/virt/if_octets-vnet0 -&gt; rate<br>
                  a0/virt/if_packets-vnet0 -&gt; rate<br>
                  a0/virt/memory-actual_balloon -&gt; absolute<br>
                  a0/virt/memory-rss -&gt; absolute<br>
                  a0/virt/memory-total -&gt; absolute<br>
                  a0/virt/ps_cputime -&gt; rate<br>
                  a0/virt/total_requests-flush-h<wbr>dc -&gt;  rate<br>
                  a0/virt/total_requests-flush-v<wbr>da<br>
                  a0/virt/total_time_in_ms-flush<wbr>-hdc -&gt; rate<br>
                  a0/virt/total_time_in_ms-flush<wbr>-vda<br>
                  a0/virt/virt_cpu_total -&gt; rate<br>
                  a0/virt/virt_vcpu-0 -&gt; rate<br>
                  a0/virt/virt_vcpu-1<br>
                <br>
                collectd &quot;just&quot; reports the changes since the last
                sampling. I&#39;m not sure which is the best way to handle
                that; I&#39;ve sent a mail to collectd list some time ago,
                no answer so far.</div>
            </blockquote>
            <div><br>
            </div>
            <div>Can you CC on that thread?</div>
            <div>I don&#39;t know how ES would work with rates at all. </div>
            <div>I want to be able to show CPU usage over time and I
              need to know if its 80% or 10%.</div>
            <br>
          </div>
        </div>
      </div>
    </blockquote>
    <br></div></div>
    Thanks to the awkward gmail interface I can&#39;t reply to myself and CC
    other people, but I can share the link:<br>
    <br>
<a class="m_-7423557398579924391moz-txt-link-freetext" href="https://mailman.verplant.org/pipermail/collectd/2017-January/006965.html" target="_blank">https://mailman.verplant.org/<wbr>pipermail/collectd/2017-<wbr>January/006965.html</a><span class=""><br>
    <br>
    <pre class="m_-7423557398579924391moz-signature" cols="72">-- 
Francesco Romani
Red Hat Engineering Virtualization R &amp; D
IRC: fromani</pre>
  </span></div>

</blockquote></div><br></div>