Cilium Wars: Scaling Up Your Network with Jedi-like Cilium Handling Techniques - Yash Tandon
summary
Shares concrete scaling outcomes from running Cilium on multi-tenant clusters with >3 years of ops experience and a real case where defaults choked a cluster of 500–600 microservices causing latency and resource blowups from sidecars (example: 500 services × 4 pods each). Describes using eBPF-powered datapath, the Cilium operator and CLI for debugging, Hubble UI visibility plus Prometheus + Grafana for alerts, and prescribes precise, label-driven CNPS, testing, and HA/load‑balancing patterns that actually reduced incidents.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 0:38Introduction to Yash
- 1:11Yoda
- 1:27Agenda
- 1:50Challenges in scaling
- 2:22Latency
- 4:11Importance of scaling
- 4:52Cost per resource
- 5:49Jedi
- 6:06Why Cilium
- 7:43Core Functions
- 9:03Architecture
- 10:44Cilium Example
- 11:24Cilium Debugging
- 12:10Monitoring Tools
- 12:53Summary
- 13:55Questions
related talks
20mTackling Container Networking at Scale
05mLightning Talk: Unveiling 5 Essential Insights from Cilium Implementation in a...
25mThat’s Just My Cup of Tea: Configuring Cilium for Performance and Scale
28mConnecting the World: Your Hands-on Guide To Cilium ClusterMesh and MCS-API
24mHigh-Scale Networking for ML Workloads With Cilium
25mMore Churn No Problem: Lessons Learned Running Cilium in Production
07mConnecting Cloud Native: Cilium + eBPF Day
01h 13mCilium Workshop