Advanced Kubernetes: Lesson Learned From Building..
summary
Builds a multi-tenant Kubernetes service using a single cluster with per-customer namespaces, network policies that block metadata APIs, and nodePort for exposing custom TCP services instead of one cloud load balancer per service (avoids the cost of ~100 load balancers). Uses Prometheus operator with central Prometheus federation and Grafana dashboards for alerts, drives autoscaling of Hazelcast clusters when memory exceeds 80% memory threshold, and adds Istio for APM/service-mesh tracing, canary and blue‑green deploys to stage rollouts.
AI-written summary · about
chapters (12)
- 0:00Start
- 2:55Isolation
- 3:48Create Separate Kubernetes Clusters for Different Environments
- 5:25Dedicated Namespaces
- 17:11Kubernetes Capacity Planning Dashboard
- 19:25Prometheus Federation
- 25:41Best Practice and the Micro Service Architecture
- 29:40Swagger Code Gen
- 33:46Deployment Strategies
- 36:48Logging
- 37:52Apm
- 41:15Questions
related talks
05mLightning Talk: What Makes a Good Multi-tenant Kubernetes Solution?
38mBuilding the Perfect Infrastructure with Kubernetes
59mKubernetes - going beyond the basics
35mKubernetes journey
01h 00mArchitecting multi-tenant PaaS offerings
35mKubernetes, Lessons Learnt
36mOptimizing Resource Usage in Kubernetes
29mNavigating the Multitenancy Maze: Avoiding Kubernetes Antipattern Pitfalls