Running Kubernetes in a Manufacturing Line
summary
Built a self‑service Kubernetes platform spanning 30 data centers with a centralized management plane in Azure and local "service" clusters per site that host DNS, registry and S3‑like caches for offline failover. Uses KKP, ArgoCD, infrastructure as code and CI pipelines, with firewall object groups tracked in a spreadsheet then automated later. Delivered an MVP in 1.5 months with ~2→3.5 engineers, met the contractual 99.96% SLO (70 min/month → reached 100%), and plans edge expansion for robots and assembly lines.
AI-written summary · about
chapters (30)
- 0:00Introduction and speakers
- 1:31Industry evolution and Industry 4.0
- 4:21Choosing Kubernetes for manufacturing
- 5:47Key project targets and challenges
- 7:52Independence and high availability goals
- 8:59Organizational collaboration model
- 13:17Initial fears and project start
- 14:18Firewall automation with object groups
- 16:03Infrastructure as code and documentation
- 16:44Standardization and scaling architecture
- 17:47Architecture design with KKP
- 20:57MVP approach and first challenges
- 22:03Network setup and data center access
- 23:08Backup and disaster recovery planning
- 23:50Authentication and service management
- 24:55Pipeline evolution and improvements
- 25:38Replica-based service clusters
- 26:20Cluster management simplicity
- 26:41Ops tooling and monitoring
- 27:26Application migration and autoscaling
- 28:50Summary and future steps
- 30:38Q&A session opening
- 31:19Edge devices and robot dogs
- 32:00DNS load and capacity management
- 33:02Uptime and cluster updates
- 34:05Global manufacturing locations
- 34:49Application operator templates
- 35:11Cluster creation flexibility
- 35:55Small clusters vs single cluster
- 37:06Closing remarks
related talks
32mRunning Kubernetes in a Manufacturing Line – What Could Possibly...
41mKubernetes — A Year in Production
29mOrchestrating 10,000+ Kubernetes deployments with 1 pipeline
35mKubernetes journey
45mLearning How to Pronounce Kubernetes to Production in 3 Months!
40mFrom dev to prod. One Year of Using Kubernetes in Production
40mThe Good, the Bad, and the Ugly of Managing Production Systems with Kubernetes
47mModern continous delivery on Azure Kubernetes Service