Make load-balancing great again with Træfik
summary
Træfik is a dynamic Go-based reverse proxy that watches orchestrators (Docker Swarm, Mesos, Kubernetes) and updates routes without restarting to provide live load balancing with dynamic weights and circuit breakers. It supports WebSocket and HTTP/2, integrates automated TLS via Let’s Encrypt, exports Prometheus metrics, is available as an official Docker image, and is used in production (e.g., Mantl/Cisco) with substantial project adoption and funding.
AI-written summary · about
chapters (23)
- 0:00Introduction
- 1:04New reverse proxy
- 2:29Microservices need
- 4:19Container orchestrators
- 8:34Dynamic service discovery
- 12:29Træfik origins goals
- 14:17Træfik in Go
- 16:44Key features
- 19:11Demo with Docker
- 24:15Deploy on Kubernetes
- 30:08Secure with Let's Encrypt
- 32:14Docker image benefits
- 33:17Træfik in Mantl
- 34:03Cluster management versions
- 35:26Authentication features
- 37:13Prometheus monitoring
- 39:28Grafana monitoring demo
- 40:32Service health
- 41:23Bug reporting community
- 42:10Version 1.3 features
- 44:17Project status adoption
- 45:21Containous creation funding
- 45:42Conclusion Q&A
related talks
17mScale anything with haproxy
15mScalable load balancing and security on AWS with HAProxy Fusion
21mStateless Datacenter Load-balancing with Beamer
33mHold My Beer - Load Testing. In Production. On Autopilot.
20mOperating High Traffic Websites on Kubernetes
34mHow CNET (And Friends) Use the CNCF Landscape To Run High Traffic, Dynamic, Scalable, And Cost-Effective Websites.
21mHandling Traffic Spikes Through Automated Prioritized Load Shedding
42mPredictive Load-Balancing: Unfair but Faster & more Robust