Linux Systems Performance - Systems Performance in 50 mins - Agenda

Aug 26, 2026

Upload by Cipher

Linux Systems Performance material by Brendan Gregg, presenting a structured 50-minute discussion of six core facets of Linux performance: observability, methodologies, benchmarking, profiling, tracing, and tuning. The content targets DBAs, developers, and operations, focusing on practical ways to measure and interpret system behavior using common observability tools. Examples include load averages via uptime, CPU and disk demand considerations, moving average time constants, and quick command-line inspection through top/htop and /proc metrics.

LanguageEnglish

CategoryTechnology

FormatPDF

Pages72

Reading Time3 h 1 min

AccessFree to download

Agenda

Observability

Linux Observability Tools

Apr, 2016 Linux Systems Performance Brendan Gregg Senior Performance Architect Systems Performance in 50 mins Agenda A brief discussion of 6 facets of Linux performance: 1. Observability 2. Methodologies 3. Benchmarking 4. Profiling 5. Tracing 6. Tuning Audience: Everyone (DBAs, developers, operations,…) 1. Observability How do you measure these? Linux Observability Tools Observability Tools • Tools showcase common metrics – Learning Linux tools is useful even if you never use them: the same metrics are in GUIs • We usually use these metrics via: – Netflix Atlas: cloud-wide monitoring – Netflix Vector: instance analysis • Linux has many tools – Plus many extra kernel sources of data that lack tools, are harder to use, and are practically undocumented • Some tool examples… upGme • One way to print load averages: $ uptime 07:42:06 up 8:16, 1 user, load average: 2.27, 2.84, 2.91 • A measure of resource demand: CPUs + disks – Other OSes only show CPUs: easier to interpret • Exponentially-damped moving averages • Time constants of 1, 5, and 15 minutes – Historic trend without the line graph • Load > \# of CPUs, may mean CPU saturation – Don’t spend more than 5 seconds studying these top (or htop) • System and per-process interval summary: $ top-18:50:26 up 7:43, 1 user, load average: 4.11, 4.91, 5.22 Tasks: 209 total, 1 running, 206 sleeping, 0 stopped, 2 zombie Cpu(s): 47 .1%us, 4.0%sy, 0.0%ni, 48.4%id, 0.0%wa, 0.0%hi, 0.3%si, 0 .2%st Mem : 70197156k total, 44831072k used, 25366084k free, 36360k buffers Swap: PID 5738 1386 1 2 […] 0k USERapiprod apiprod root root total, PR NI 20 0 20 0 20 0 20 0 VIRT 62.6g 17452 24340 0 0k used, RES 29g 1388 2272 0 SHR 352m 964 1340 0 S S R S S %CPU 417 0 0 0 0k free, 11873356k cached %MEM 44.2 0.0 0.0 0.0 TIME+ 2144:15 0:00 .02 0:01 .51 0:00 .00 COMMAND java top init kthreadd • %CPU is summed across all CPUs • Can miss short-lived processes (atop won’t) • Can consume noticeable CPU to read / proc

FAQs

What six facets of Linux performance are covered in the agenda?

The agenda covers observability, methodologies, benchmarking, profiling, tracing, and tuning.

How can load averages be printed on Linux?

Use the uptime command to display load averages alongside current system status.

What does the top/htop output help you understand?

top/htop provide an interval summary for the system and each process, including CPU usage breakdown and memory usage, helping spot how load relates to resource demand.

Aug 26, 2026

Upload by Cipher