Acing Distributed Container Management
summary
An AIOps-driven platform called Cordant acts as a mothership cluster built on ClusterAPI with templated YAMLs to avoid repetitive cluster specs, Sveltos for add‑on state management, and Cosmotron for hosted control planes; observability uses Grafana, VictoriaMetrics and OpenCost to deliver predictive scaling and automated root‑cause suggestions. Demo claims thousands‑of‑cluster scale but notes production runs are still a case study; presenters warn agent sprawl (Nvidia target: 100M agents, ~2,000 per person) and cite McKinsey’s 30–50% cost reduction projection.
AI-written summary · about
chapters (17)
- 0:00Welcome
- 1:25Developer expectations
- 3:10Kubernetes control plane
- 5:16Platform engineering product
- 7:45Platform engineering challenges
- 10:32Resolving tool sprawl
- 11:54Lightweight clusters
- 13:20Cordant cluster management
- 16:14AI ops integration
- 19:48AI ops visibility problems
- 21:56AI ops evolution
- 24:08Agent sprawl problem
- 25:54Cordant agent governance
- 28:45Agentic AI ROI and risks
- 30:54Future agent capabilities
- 32:00Next steps
- 34:11Q&A
related talks
18mA Cloud-Native Tool for the mass
23mScale GitOps With the Argo CD Agent and Open Cluster Management
32mDon't Start With 500 Clusters: Patterns in Scaling Multi-Cluster Kubernetes…
29mMulticluster Kubernetes Management Made Easy With Open Cluster Management
09mCODECO: AI-Driven Orchestration for Multi-Cluster Edge Deployment
21mMulti Cluster Magics With Argo CD and Cluster Inventory
33mCentralizing Kubernetes and Container Operations
29mTowards Sustainable Multi-Cluster Management in Real-World Academic Infrastructure