Performance Bottlenecks in RHEL ?

[Chetan_Tiwary_]

What tool do you use to diagnose and resolve performance bottlenecks in RHEL ?

[Trevor]

I'll be the leadoff hitter on this one.

Since bottlenecks are going to occur with the resources, I've
got to put my hands on tools that will give me a look at CPU
usage, memory usage, storage I/O, and network I/O. This will
require a team effort to cover all the bases.

Here's my list, and the resources that they'll assist with in the
discovery of bottlenecks:

- top -- will provide a real-time report of system resource
usage that includes, CPU usage, memory usage, and running
processes. It certainly makes sense to want to keep an eye
on the processes since so much of what's happening in a Linux
environment involves processes.

- iostat -- displays the load average of the CPU, along with
disk I/O information. This tool will provide a look at both
local and remote disks (e.g. NFS) activity.

- vmstat -- as the man page states, it reports informaiton about
processes, memory, paging, block I/O, disks activity, and cpu
activity. We definitely want to keep an eye on paging stats - this
can be a bottleneck like no other.

- free -- the old standby that checks memory usage.

- netstat -- a tool that's going to give us a look at network
statistics. Definitely a go-to tool when network performance
needs to be examined.

- iotop -- a cousin of the top command, is worthy of consideration,
considering it provides real-time information on Disk I/O and
processes. Being able to have a look at the exact process that's
responsible for a lot of disk read/write activity - which is directly
related to disk I/O - is very valuable.

- sar -- for me, this is the equivalent of a baseball player batting
in the cleanup spot of the lineup. this all-in-one tool can be used
to give an close-up look at CPU usage, Memory usage, I/O devices
consumption, Network monitoring, Disk usage, process and thread
allocation, and even information of the filesystems that are mounted.
This tool provides a lot of support to the others - or is it the other
way around

There are other tools that are available, but this list represents my
starting lineup!!!

[niloy_kumar]

in addition to the above we use new relic integration for linux and kubernetes add-ons to monitor pods

[nikijain]

top, vmstat, netstate

[Kent-Kamau]

I use top to help me check resource usage and manage performance bottlenecks.

[juicylemon]

top - current cpu and memory
ps - running processes
free -m  - memory consumption
uptime - load view 

[Aamir_Ahmed]

my way to diagnose the performance bottleneck woluld be start with 'df'.=> top (here you'll get the idea about usage of resource, and accordingly you can check the stats ie. iostat/vmstat/nsfstat) => sar => dmesg/journelctl

[Matt-Harding]

sar, sar, sar - always sar! Hasn't changed in... 40+ years?

Also, nmon for a nice all-in-one GUI that incorporates a lot of stuff from iostat, top etc in one place.

Long live Linux.

[Trevor]

Let's hear for sar!!!!!

[mcapileibm]

This helps me diagnose performance bottlenecks by identifying CPU saturation, memory pressure, or excessive I/O wait times