CPU usage [Metricbeat Apache] No apache.status.cpu.load

Hi community.

I'm trying to get all metrics of my apache through metricbeat but I'm not getting apace.status.cpu.* metrics

I've been googling about this possible issue, but I've not find any information. Does any one get the same issue fixed?

elasticsearch version: 6.3.2
metricbeat version: 6.3.2
S.O Centos 7
Apache Version: 2.4

1 Like

I have the exact same issue on the save versions with the exception that i am running Ubuntu 14.04.

I am missing:
etc. only seeing scoreboard etc.

here is an example of my apache server-status:

<title>Apache Status</title>
<h1>Apache Server Status for localhost (via</h1>

<dl><dt>Server Version: Apache/2.4.7 (Ubuntu) PHP/5.5.9-1ubuntu4.25</dt>
<dt>Server MPM: prefork</dt>
<dt>Server Built: Apr 18 2018 15:36:26
</dt></dl><hr /><dl>
<dt>Current Time: Thursday, 13-Sep-2018 18:46:44 UTC</dt>
<dt>Restart Time: Thursday, 13-Sep-2018 18:26:58 UTC</dt>
<dt>Parent Server Config. Generation: 1</dt>
<dt>Parent Server MPM Generation: 0</dt>
<dt>Server uptime:  19 minutes 45 seconds</dt>
<dt>Server load: 0.00 0.03 0.05</dt>
<dt>Total accesses: 157 - Total Traffic: 85 kB</dt>
<dt>CPU Usage: u.02 s.04 cu0 cs0 - .00506% CPU load</dt>
<dt>.132 requests/sec - 73 B/second - 554 B/request</dt>
<dt>2 requests currently being processed, 4 idle workers</dt>
<p>Scoreboard Key:<br />
"<b><code>_</code></b>" Waiting for Connection,
"<b><code>S</code></b>" Starting up,
"<b><code>R</code></b>" Reading Request,<br />
"<b><code>W</code></b>" Sending Reply,
"<b><code>K</code></b>" Keepalive (read),
"<b><code>D</code></b>" DNS Lookup,<br />
"<b><code>C</code></b>" Closing connection,
"<b><code>L</code></b>" Logging,
"<b><code>G</code></b>" Gracefully finishing,<br />
"<b><code>I</code></b>" Idle cleanup of worker,
"<b><code>.</code></b>" Open slot with no current process<br />
<p />

<table border="0"><tr><th>Srv</th><th>PID</th><th>Acc</th><th>M</th><th>CPU

</td><td>localhost</td><td nowrap></td><td nowrap>NULL</td></tr>

</td><td>localhost</td><td nowrap></td><td nowrap>GET /server-status?auto= HTTP/1.1</td></tr>

</td><td>localhost</td><td nowrap></td><td nowrap>NULL</td></tr>

</td><td></td><td nowrap></td><td nowrap>GET /server-status HTTP/1.1</td></tr>

</td><td></td><td nowrap>webship.test300.rsis.io:80</td><td nowrap>GET /ping.php HTTP/1.1</td></tr>

</td><td>localhost</td><td nowrap></td><td nowrap>NULL</td></tr>

 <hr /> <table>
 <tr><th>Srv</th><td>Child Server number - generation</td></tr>
 <tr><th>PID</th><td>OS process ID</td></tr>
 <tr><th>Acc</th><td>Number of accesses this connection / this child / this slot</td></tr>
 <tr><th>M</th><td>Mode of operation</td></tr>
<tr><th>CPU</th><td>CPU usage, number of seconds</td></tr>
<tr><th>SS</th><td>Seconds since beginning of most recent request</td></tr>
 <tr><th>Req</th><td>Milliseconds required to process most recent request</td></tr>
 <tr><th>Conn</th><td>Kilobytes transferred this connection</td></tr>
 <tr><th>Child</th><td>Megabytes transferred this child</td></tr>
 <tr><th>Slot</th><td>Total megabytes transferred this slot</td></tr>
<hr />
<address>Apache/2.4.7 (Ubuntu) Server at localhost Port 81</address>

Noticed same issue on CentOS 7. No cpu fields.. With 'ExtendedStatus On'.

Hi all :slight_smile:

I have just done a quick check using the latest Metricbeat version (6.4) on Arch without problems. Can you try with this version, please? Maybe it's something specific of RH based distros but I'd like to discard a version issue.


Hi again,

We have checked that in Centos 7 the version of the server that is installed by default is 2.4.6 and I'm afraid that the CPU load wasn't included in this version yet.

Anyways, you can extract this data by using the proccess metricset which will give you the CPU use of the server proccess too.

I hope it helps

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.