We thought that the world right now does not need another open source tool that gathers and transports system metrics (there are already so many: collectd (https://collectd.org/) is a very stable and proven technology to gather system information. collectd provides a bunch of plugins (https://collectd.org/related.shtml). What the world is missing though is an integrated open source APM solution that provides in-depth application monitoring and analytics combined with the base metrics of the hardware. inspectIT does provide the application monitoring. For the system and OS part, we rely on collectd. collectd (wiki/index.php/Table_of_Plugins) to gather information.
Requirements
- Root rights to install collectd
- *nix system
- Connection to the longterm storage is possible - default port is 8096
Configuration
This section describes how you setup collectd and send the data to our longterm storage.
- Installation of collectd
collectd is often provided by the package manager of *nix based systems. If your system is not supported, please find collectd at their homepage at https://collectd.org/
...
What we provide is to link all of it together. We allow the collectd agent (we just call it HostAgent) to write its data into our monitoring database. The data can then be used as basis for everything else. We are currently building top-notch anomaly detection on top of the data within our monitoring database. So guess what, your system metric information will be part of it! You can show the system information next to your application information, next to the business/usecase information. In a single dashboard. A single pane of glass.
This is what we want to achieve!
Configurations
...
- download.shtml. For windows-based servers you might want to look at http://ssc-serv.com/ (which supports collectd, but is commercial). Please understand that this description is for the *nix based setup with collectd. Please contact us for question to windows directly.
- Allow that data collection within the longterm data storage
Please verify that you configured the long term data storage to receive data from our collectd Host Integration. For a detailed explanation please see Allow data collection with inspectIT collectd Integration Configure to send the data to the longterm storage
Please find the file /etc/collectd.conf and configure the section accordingly:
Code Block Hostname "inspectIT-collectd" #FQDNLookup true BaseDir "/var/lib/collectd" PIDFile "/var/run/collectd.pid" PluginDir "/usr/lib64/collectd" TypesDB "/usr/share/collectd/types.db"- Scroll down to the LoadPlugin section and uncomment the LoadPlugin network and LoadPlugin uptime lines (that is remove the # at the beginning of the line)
Scroll down to the Plugin Configuration setting. Now we need to specify where the data should be send. Please add the new network plugin here accordingly:
Code Block . . . ############################################################################## # Plugin configuration # #----------------------------------------------------------------------------# # In this section configuration stubs for each plugin are provided. A desc- # # ription of those options is available in the collectd.conf(5) manual page. # ############################################################################## <Plugin network> Server "127.0.0.1" "8096" </Plugin> #<Plugin "aggregation"> . . .
Verify configuration
To verify the configuration please execute
| Code Block |
|---|
sudo /usr/sbin/collectd -t |
If no errors are present this command will return with no output. Else you get a detailed explanation of the errors. Please note that this test is just verifying that the configuration on the collectd daemon is correct. It does not actually test the path to the longterm storage.
Startup
You can start the daemon by running sudo service collectd start