Optimizing costs across several OpenShift clusters
summary
Concrete steps and measured results: the team uses Prometheus and Grafana plus OC and cloud cost explorers to profile usage, then applies fixes—scaling workers to zero, switching Intel→AMD machine types, using heterogeneous clusters with ARM and Hypershift, enabling autoscaler rules and node down‑pressure, switching storage to gp3, committing to Savings Plans/reserved instances, and running spot instances via Spot.io. A pull‑through registry cache to S3 and collocating data cut egress, NAT gateway removal lowers network bills, and the big batch of changes yields about 60% peak savings.
AI-written summary · about
chapters (10)
- 0:00Optimizing OpenShift costs
- 1:29Infrastructure cost analysis
- 2:56Quick cost reduction wins
- 6:15Machine type optimization
- 9:06Efficient cluster scaling
- 12:18Spot instances for savings
- 14:19Data transfer cost reduction
- 18:38Measuring optimization impact
- 23:20When to optimize costs
- 29:00Q&A
related talks
26mDescubra las tendencias de optimización de costos en Kubernetes
57mKubernetes Cloud Cost Monitoring with OpenCost & Optimization Strategies
15mDemo: Resource(Cost) optimization on planet scale Kubernetes clusters.
13mOptimizing Workload Management in a Multi-Cluster Environment: A Case Study
44mFinOps at scale: Lessons from scaling Kubernetes workloads
34mOpen Source Cloud Costs: A Deep Dive Into OpenCost's Impact on Enterprise...
36mOptimizing Resource Usage in Kubernetes
39mOptimizing Resource Usage in Kubernetes