3

I have setup a monitoring server with the following setup.

<Plugin network>
Listen "0.0.0.0" "25826"
</Plugin>

Now my clients are sending data to the monitoring server(verified through tcpdump). Even the collection folder shows that the data is being dumped

/var/lib/collectd/rrd
[ec2-user at x rrd]$ ll
total 4
drwxr-xr-x 11 root root 4096 Nov 20 17:53 x-web-1.y.com
[ec2-user at x rrd]$

I have also verified with find . -mmin 1 to see if its being constantly updated.

[ec2-user@x rrd]$ find . -mmin 1
./x-web-1.y.com/interface-eth0/if_errors.rrd
./x-web-1.y.com/interface-eth0/if_packets.rrd
./x-web-1.y.com/interface-eth0/if_octets.rrd
./x-web-1.y.com/disk-xvda1/disk_time.rrd
./x-web-1.y.com/disk-xvda1/disk_ops.rrd
./x-web-1.y.com/disk-xvda1/disk_octets.rrd
./x-web-1.y.com/disk-xvda1/disk_merged.rrd

But when i look it up through collectd-web, I don't see the clients enter image description here

What might be wrong in my setup?

Improve this question
1
  • I am having a similar problem though my servers /var/lib/collectd folders are not populating with client data and log files are showing nothing, on either client or server. Were you able to figure out your problem?
    – Joey T
    Jan 15, 2013 at 4:40

4 Answers 4

Reset to default
0

The posted image doesn't seem a part of collectd-web. Anyway, there's a sort of cache effect with Firefox which prevents the rebuilding of collectd-web menus, included the hosts list. Clean Firefox' cache and try again.

Improve this answer
1
  • the snapshot is from the collection3 perl cgi app included with collectd source file. Is the collectd-web a different app entirely or a separate package included with some distros?
    – Joey T
    Jan 16, 2013 at 19:03
0

As commented previously, I was having the same problem using the default collectd.conf file, which defaults the global Host parameter as localhost. I discovered that if any of the clients collectors are using the same global Host name, this will confuse your Server collector which usually is evidenced by errors in the log file that too many updates are occurring for the same timestamp.

The solution is to ensure that all of your clients are using a unique global Host parameter, at that point your collecter server's rrd folder should populate with folders containing the individual client hostnames with captured sub-content.

Improve this answer
0

I was having this issue as well. It happened when I was serving collectd-web through Apache, instead of using the built-in python server (where it was working ok).

What was happening was that apache was not processing the gci files in collectd-web. To fix it you just have to make sure you have mod_gci installed, and then edit your /etc/apache2/sites-available file to allow apache to allow configuration overrides in your collectd-web folder:

<Directory /var/www/collectd-web>
  AllowOverride all
</Directory>

This will cause apache to load collectd-web's .htacccess file (on the cgi-bin folder), which contain the directives for executing the cgi scripts.

This small gotcha is mentioned on the docs: http://collectdweb.appspot.com/documentation/

Not sure if you're using apache as well, but whatever server you're using: check the cgi scripts are being evaluated. Otherwise, the dinamic parts of the interface will not load.

Improve this answer
0

Consider switching to prometheus or some other monitor system that are up to date.

At this moment collectd doesn't even have a ubuntu 22 LTS package.

We switched away from collectd to Prometheus. Much easier to handle and more options.

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .