In this talk, Jonathan Looney from Netflix describes the Open Connect Appliance and the optimizations that make it possible to serve over 180 GB/s of TLS-encrypted connections with less than 50% CPU on a single 32-core 2.5-GHz CPU in 2 RU.

The post gives an overview of where the complexities in load-testing lie.

Ethr is a golang-based network performance measurement tool similar to ip iPerf3.

Benchmark for storage workloads to predict how a storage device will perform in real production environments.

With the presented tools BTF and CO-RE, BPF gets more portable, not requiring the whole build chain with LLVM, Clang, and kernel header dependencies. In this blog post, Brendan Gregg explains how it works and what it means for BPF performance tools.

Canonical introduces a profiling tool called etrace. While being designed as a generic tracing application, it is specially tailored to analyze snaps.

BPF originally stood for Berkeley Packet Filter, but has been extended in Linux to become a generic kernel execution engine, capable of running a new type of user-defined and kernel-mode applications.

This is the official site for the book BPF Performance Tools: Linux System and Application Observability, published by Addison Wesley (2019)

Today’s software systems are arguably robust at logging and recovering from fail-stop hardware – there is a clear,binary signal that is fairly easy to recognize a and interpret. We believe fail-slow hardware is a fundamentally harder problem to solve. It is very hard to distinguish such cases from ones that are caused by software performance issues. It is also evident that many modern,advanced deployed systems do not anticipate this failure mode. We hope that our study can influence vendors, operators, and systems designers to treat fail-slow hardware as a separate class of failures and start addressing them more robustly in future systems.

This post explains Transparent Hugepages (THP) in a nutshell, describes techniques that can be used to measure the performance impact, shows the effect on a real-world application.

Easy flamegraphs for Rust projects and everything else, without Perl or pipes.

Conclusions:

1. Network performance and utilization will affect the general application throughput.
2. Check if you are hitting network bandwidth limits
3. Protocol compression can improve the results if you are limited by network bandwidth, but also can make things worse if you are not
4. SSL encryption has some penalty (~10%) with a low amount of threads, but it does not scale for high concurrency workloads.

Critical but oft-neglected service metrics that every SRE and product owner should care about.

Resource pressure metrics from the Linux kernel.

Gain insight into resource utilization with new Linux kernel pressure metrics and related tools.