The Netflix TechBlog

OCTOBER 19, 2020

In our previous blog post we introduced Edgar, our troubleshooting tool for streaming sessions. If we had an ID for each streaming session then distributed tracing could easily reconstruct session failure by providing service topology, retry and error tags, and latency measurements for all service calls.

Infrastructure 233

Infrastructure Transportation Storage Open Source 233

Adrian Cockcroft

JANUARY 29, 2023

There is no way to model how much more traffic you can send to that system before it exceeds it’s SLA. Bill Kaiser of NewRelic published this blog in 2017 which goes some way towards what I’m talking about, but since then I have figured out a new way to interpret the data. Mu is the mean of each component, the latency.

Lambda 98

Lambda Latency Cache C++ 98

The Symphonia

MAY 22, 2018

Gojko Adzic has done some great speaking and writing on his experience here, and I included the link to his talk from late 2017. I also rewrote the section on Startup Latency since Cold Starts are one of the big “FUD” areas of Serverless. I was glad to be able to talk about Amazon’s automated traffic shifting / canary releases.

Serverless 40

Serverless Architecture Lambda Azure 40

Brendan Gregg

SEPTEMBER 25, 2021

A Cassandra database cluster had switched to Ubuntu and noticed write latency increased by over 30%. Since instances of both CentOS and Ubuntu were running in parallel, I could collect flame graphs at the same time (same time-of-day traffic mix) and compare them side by side. You have two hands: observation and experimentation.

Speed 126

Speed Java AWS Virtualization 126

Brendan Gregg

SEPTEMBER 5, 2017

Here's some output from my zfsdist tool, in bcc/BPF, which measures ZFS latency as a histogram on Linux: # zfsdist. Tracing ZFS operation latency. Search for recent blog posts on Linux containers, and try them out, and you'll piece together their capabilities and workings bit by bit. Hit Ctrl-C to end. ^C

Virtualization 76

Virtualization AWS Engineering Hardware 76

Smashing Magazine

AUGUST 10, 2023

DebugBear explains this nicely in its blog: Simulated throttling provides low variability and makes test quick and cheap to run. In contrast, tools like DebugBear and WebPageTest use more realistic throttling that accurately reflects network round trips on a higher-latency connection. Real usage data would be better, of course.

Speed 81

Speed Monitoring Testing Network 81

Brendan Gregg

SEPTEMBER 25, 2021

A Cassandra database cluster had switched to Ubuntu and noticed write latency increased by over 30%. Since instances of both CentOS and Ubuntu were running in parallel, I could collect flame graphs at the same time (same time-of-day traffic mix) and compare them side by side. You have two hands: observation and experimentation.

Speed 40

Speed Java AWS Virtualization 40