How to install and use Performance Co-Pilot

How to install and use Performance Co-Pilot


Here in this article we will see how we can install Performance Co-Pilot (PCP) for collecting performance metrics for local and generate performance charts on the data.

Test Environment

Fedora 35 workstation

What is Performance Co-Pilot (PCP)

Performance Co-Pilot provides a framework and services to support system level performance monitoring and management. It presents a layer of abstraction to the system for collecting the performance data. It provides a set of tools for collecting and processing the performance data.


Step1: Create a repository

As a first step create the below repository file which is required for installing the package.

[admin@fedser ~]$ cat /etc/yum.repos.d/performancecopilot.repo
name=Performance Co-Pilot

Step2: Install Performance Co-Pilot (PCP)

[admin@fedser ~]$ sudo dnf install pcp-zeroconf pcp-gui
  pcp-5.3.7-1.fc35.x86_64                pcp-conf-5.3.7-1.fc35.x86_64             pcp-doc-5.3.7-1.fc35.noarch                pcp-libs-5.3.7-1.fc35.x86_64     
  pcp-pmda-dm-5.3.7-1.fc35.x86_64        pcp-pmda-nfsclient-5.3.7-1.fc35.x86_64   pcp-pmda-openmetrics-5.3.7-1.fc35.x86_64   pcp-selinux-5.3.7-1.fc35.x86_64  
  pcp-system-tools-5.3.7-1.fc35.x86_64   pcp-zeroconf-5.3.7-1.fc35.x86_64         python3-pcp-5.3.7-1.fc35.x86_64           


Step3: Start the Performance metrics collection service

Before the tool can start collecting the performance metrics we need to make sure the performance metrics collection daemon service is up and running as shown below.

[admin@fedser ~]$ sudo systemctl start pmcd.service 
[admin@fedser ~]$ sudo systemctl status pmcd.service 
● pmcd.service - Performance Metrics Collector Daemon
     Loaded: loaded (/usr/lib/systemd/system/pmcd.service; enabled; vendor preset: enabled)
     Active: active (running) since Tue 2022-05-31 15:30:55 IST; 1min 26s ago
       Docs: man:pmcd(1)
   Main PID: 65943 (pmcd)
      Tasks: 9 (limit: 18871)
     Memory: 40.3M
        CPU: 2.601s

Step4: List the available performance metrics

Once the collector service is started we can list the performance metrics that can be collected from the local system.

[admin@fedser ~]$ pminfo

To get a brief information about what a particular metrics collects we can run the below.

[admin@fedser ~]$ pminfo -t kernel.all.uptime
kernel.all.uptime [time the current kernel has been running]

Step5: Record metrics on local system

[admin@fedser ~]$ sudo systemctl start pmlogger.service 
[admin@fedser ~]$ sudo systemctl status pmlogger.service 
● pmlogger.service - Performance Metrics Archive Logger
     Loaded: loaded (/usr/lib/systemd/system/pmlogger.service; enabled; vendor preset: enabled)
     Active: active (running) since Tue 2022-05-31 15:42:38 IST; 2s ago
       Docs: man:pmlogger(1)
   Main PID: 67877 (pmlogger)
      Tasks: 1 (limit: 18871)
     Memory: 3.3M
        CPU: 573ms

Verify that the performance metrics logger instance is running as shown below.

[admin@fedser ~]$ pcp
Performance Co-Pilot configuration on

 platform: Linux 5.17.7-200.fc35.x86_64 #1 SMP PREEMPT Thu May 12 14:56:48 UTC 2022 x86_64
 hardware: 8 cpus, 1 disk, 1 node, 15770MB RAM
 timezone: IST-5:30
 services: pmcd
     pmcd: Version 5.3.7-1, 12 agents, 1 client
     pmda: root pmcd proc pmproxy xfs linux nfsclient mmv kvm jbd2
           dm openmetrics
 pmlogger: primary logger: /var/log/pcp/pmlogger/

Verify that the archive files are getting created as shown below.

[admin@fedser ~]$ ls -ltr /var/log/pcp/pmlogger/
total 3388
-rw-r--r--. 1 pcp pcp     250 May 31 15:42 Latest
-rw-r--r--. 1 pcp pcp   27603 May 31 15:42 pmlogger.log
-rw-r--r--. 1 pcp pcp  339885 May 31 15:49 20220531.15.42.meta
-rw-r--r--. 1 pcp pcp     712 May 31 15:49 20220531.15.42.index
-rw-r--r--. 1 pcp pcp 3087192 May 31 15:49 20220531.15.42.0

Step6: Generate charts on performance metrics

As a first step we need to launch the below GUI tool for viewing the charts for a collection of metrics and their performance.

[admin@fedser ~]$ pmchart

Once the GUI is loaded we can go into File – Open View which will show some default charts views available for us to view. We can select any one of them and view the performance metrics data live on the GUI as shown below.

Hope you enjoyed reading this article. Thank you..