Performance Co-Pilot

Performance Co-Pilot is a system performance analysis toolkit.

Collect

  • Lightweight : Collect performance metrics from your systems efficiently.
  • Distributed : Collate metrics from multiple hosts and a variety of operating systems.
  • Included : Everything you need is already included in the major distributions: Fedora, RHEL, Debian, SUSE, Ubuntu, Gentoo

Analyze

Analyze systems' performance metrics in real-time or using historical data.

Compare performance metrics between different hosts and different intervals. Observe trends and identify abnormal patterns.

Extend

Extend the collected performance metrics in a simple way.

PCP offers a multitude of APIs and libraries to extract and make use of performance metrics from your own application.

Use Performance Co-Pilot Now

News

Oct 2021
PCP 5.3.4

Bug fix release, especially in systemd integration for pmlogger farms. Adds new detailed processor frequency scaling metrics on Linux. Improvements to the pcp-atop and pcp-htop utilities.

Sep 2021
PCP 5.3.3

New eBPF ELF binary PMDA based on libbpf. Improved pmdabcc memory footprint. Dynamic columns in pcp-htop allowing columns for any process metric in PCP. New Linux autogroup metrics. New and improved processor and block device saturation pmie rules.

Jul 2021
PCP 5.3.2

New socket statistics utility - pcp-ss . New metric collector - denki denki (power), and updates to sockets, ds389, HA cluster and Linux kernel metrics. Further improvements to threading and memory handling in the pmproxy archive discovery services and REST API.

Jun 2021
PCP 5.3.1

Minor bug fix release. Major performance optimizations for the pmproxy REST APIs, archive discovery and time series queries, as well as for the Linux proc metrics. Fixes and new configuration options for the pmrep utility.

Apr 2021
PCP 5.3.0

Scalability and memory footprint improvements in the pmproxy service. pcp-atop allows process filtering based on state, sorting threads per process and additional reporting features (kernel wait channel symbol for blocked processes, VM OOM process killing, swap-cached, KSM, ZFS and zswap memory metrics, per-process locked memory and scheduler run delay metrics). pcp-htop added the SysArch Meter and ZFS ARC statistics. pmrep allows specifying instances with PID. New Linux kernel metrics for fibre channel HBAs, pseudo TTYs and much improved interrupts metric handling.

Feb 2021
PCP 5.2.5

New Linux ZFS, High Availability (Pacemaker) and detailed sockets metrics. pmproxy memory management improvements.

Dec 2020
PCP 5.2.3

Performance improvements and process network statistics support in the pcp-atop command. Experimental pcp-htop command added. Improvements to the pmproxy REST API for fast, scalable pmseries querying.