Intro + Deep Dive: SIG Scalability

on YouTube ↗39m

" summary

SIG Scalability defines a multi‑dimensional scalability envelope and enforces concrete SLOs (e.g., 99th‑percentile mutating API latency ≤1s in 5‑minute windows, max two windows missed per day), runs automated tests with cluster-loader2 and daily 5k‑node simulations via Cubemark hollow nodes (80 real nodes simulating 5k), keeps a 100‑node presubmit performance job, uses perfdash and CPU/memory profiles to find regressions (Go compiler and priority/fairness fixes) and pushes specific fixes.

AI-written summary · about

" chapters (8)

1. 0:00SIG Scalability intro
2. 3:35Kubernetes scalability
3. 9:22Scalability envelope & SLOs
4. 17:36Scalability testing tools
5. 23:26Recent scalability work
6. 25:12Join SIG Scalability
7. 36:51Proactive scalability
8. 38:17Q&A

" related talks

• 43mIntro + Deep Dive: SIG ScalabilityMarcel Zięba, Wojciech Tyczyński · KubeCon + CloudNativeCon Europe 2022
• 29mIntro + DeepDive: SIG ScalabilityMarcel Zięba, Wojciech Tyczyński · KubeCon + CloudNativeCon North America 2021
• 31mScaling a Distributed Container Management System: Diego PerformanceJames Myers, Luan Dos Santos · Cloud Foundry Summit Silicon Valley 2016
• 35mDeep Dive: Kubernetes (Instrumentation) SIGMax Leonard Inden, Frederic Branczyk · KubeCon + CloudNativeCon Europe 2019
• 37mUsing SLOs for Continuous Performance Optimizations of Your K8s WorkloadsAndreas Grabner · KubeCon + CloudNativeCon North America 2021
• 48mScale and Performance Testing of Kubernetes: Answers for Specific ApplicationsGeorgy Okrokvertskhov, Aleksandr Shaposhnikov · KubeCon + CloudNativeCon North America 2016
• 30mAbove the Clouds: Investigating Systems from 32,000 ft.Pat Eyler · Cascadia Ruby 2012
• 18mScaling APIs to Billions: Extreme-Scale ReliabilityNikhil Kokal · Conf42 DevOps 2026