Brendan Gregg
MARCH 16, 2024
2015-2020: Overhead As part of production rollout I did many performance overhead tests, which I've described publicly before: The overhead of adding frame pointers to everything (libc and Java) was usually less than 1%, with one exception of 10%. The actual overhead depends on your workload. Adding anything will cause the same effect.
Brendan Gregg
SEPTEMBER 25, 2021
Was there some other program consuming CPU, like a misbehaving Ubuntu service that wasn't in CentOS? As a Xen guest, this profile was gathered using perf(1) and the kernel's software cpu-clock soft interrupts, not the hardware NMI. I also shared setting the clocksource in my talks and in my 2015 [Linux tunables] post.
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Brendan Gregg
FEBRUARY 28, 2023
I'm now program co-chair for SREcon 2023 APAC, and our 2023 conference is June 14-16 in Singapore. And now, helping bring USENIX conferences to Australia by giving the first keynote: I could not have scripted or expected it. The call for participation ends on March 2nd 23:59 SGT!
Randon ASCII
FEBRUARY 25, 2018
In this particular investigation, which spanned twenty months, we suspected hardware failure, compiler bugs, linker bugs, and other possibilities. Jumping too quickly to blaming hardware or build tools is a classic mistake, but in this case the mistake was that we weren’t thinking big enough. failure rate. Russian translation is here.
Brendan Gregg
FEBRUARY 28, 2023
I'm now program co-chair for SREcon 2023 APAC, and our 2023 conference is June 14-16 in Singapore. And now, helping bring USENIX conferences to Australia by giving the first keynote: I could not have scripted or expected it. It was a great privilege. The call for participation ends on March 2nd 23:59 SGT!
Randon ASCII
AUGUST 16, 2018
When a Windows program stops pumping messages there will be ETW events emitted to indicate exactly where this happened , so those types of hangs are trivial to find. Normally running every 6 ms is enough to make a program responsive, but for some reason it wasn’t making any progress. But apparently Chrome kept on pumping messages.
Brendan Gregg
SEPTEMBER 25, 2021
Was there some other program consuming CPU, like a misbehaving Ubuntu service that wasn't in CentOS? As a Xen guest, this profile was gathered using perf(1) and the kernel's software cpu-clock soft interrupts, not the hardware NMI. I also shared setting the clocksource in my talks and in my 2015 [Linux tunables] post.
Expert insights. Personalized for you.
145 
Let's personalize your content