It's a fast-paced business world, and systems administrators are always kept busy. Red Hat support engineers are always on the lookout for tools that can streamline processes and deliver quick, actionable insights, especially when helping customers. Recently, I've been using the pcp-xsos
utility to help me with resource utilization analysis.
Problem: Time-consuming analysis
As a sysadmin, I often start a system analysis using detailed summaries from performance recordings of each process that's utilizing memory, CPU, IO, and so on. The Performance Co-Pilot (PCP) included with Red Hat Enterprise Linux (RHEL) is a powerful toolkit with lots of useful commands and options, but it's easy to become overwhelmed with so much data. Simply put, drawing an actionable conclusion from your performance data can be a lengthy process, especially when dealing with large PCP archive files.
Solution: pcp-xsos
The pcp-xsos
tool reduces analysis time from minutes to mere seconds. It generates high-level resource utilization summaries directly from a PCP archive, using a single-sample-any-offset approach to simplify the data review process. This drastically reduces the time you need to understand the state of your system.
Click the image below to play the screencast and see it in use:
Benefits of pcp-xsos
- Efficiency: With
pcp-xsos
, you can get a comprehensive resource summary in just seconds. This saves you valuable time during the analysis session. A rapid turnaround is particularly crucial during severe incident scenarios, where every minute counts - Improved customer experience: Fast analysis often produces a quick resolution. By reducing your turnaround time, you can provide a better overall experience for your customer
- Comprehensive insights: Despite its speed,
pcp-xsos
doesn't compromise on the depth of information it provides. It offers detailed insight into memory consumption, CPU usage, and other critical metrics, so you can get up to speed quickly and start formulating a solution - Easy to use: The
pcp-xsos
utility simplifies the otherwise complex task of sifting through multiple PCP commands and options, making it accessible to seasoned engineers and new users alike
Try pcp-xsos
You don’t need to wait to try out pcp-xsos
. It's soon going to be a built-in part of RHEL, but the tool works with any version of PCP you already have installed, and it has no additional dependencies beyond PCP. Download it and try it out for yourself:
$ sudo yum install pcp-zeroconf
$ sudo wget -o /usr/libexec/pcp/bin/pcp-xsos \ https://github.com/performancecopilot/xsos-pcp/blob/main/pcp-xsos
$ pcp xsos
It’s that easy. In no time, you have an operating system summary, reported live from your running system. Once you have some recorded data in /var/log/pcp/pmlogger
(the pcp-zeroconf package install sets that up automatically), you can use the --archive
option to get a historical view, too.
Performance Co-Pilot archives are platform-neutral, and our commitment to backwards compatibility means you always know that new tools like pcp-xsos
can replay them even if that data was recorded years ago.
Working on an embedded ARM device? Or maybe an s390 mainframe? Or an x86 laptop? Not a problem! PCP archives from any host can be replayed on any other host.
The pcp-xsos
is an easy tool that's well worth adding to your analysis toolkit. It has a clean and simple interface matched only by its sampling efficiency. Give it a try today, because you deserve a break!
关于作者
Nathan is an engineer in Red Hat's Platform Tools group, leading the Grafana and PCP team.
产品
工具
试用购买与出售
沟通
关于红帽
我们是世界领先的企业开源解决方案供应商,提供包括 Linux、云、容器和 Kubernetes。我们致力于提供经过安全强化的解决方案,从核心数据中心到网络边缘,让企业能够更轻松地跨平台和环境运营。