A DevOps State of Mind: Managing Microservices and Databases with Kubernetes Presented by: Presented by: Chris Van Tuin - Red Hat - Chief Technologist (firstname.lastname@example.org).
Rapid innovation, changing business landscapes, and new IT demands force businesses to make changes quickly. In the eyes of many, DevOps + Microservices + Containers are at the brink of becoming pervasive in IT to accelerate business innovation. In this presentation, you'll learn about managing containerized Microservices # Databases at scale: - Best practices for immutable and secure container images - Deployment strategies for Microservices including Recreate, Rolling, Blue/Green, A/B testing with Canaries Managing Database migrations in a CI/CD pipeline
# Automated Software Factory - Key to k8s can be deployed on bare metal. - Declaratively defined application deployment - Create pods. role:web, role: app - Route outbound public traffic via ingress / routes - Define a pod to monitor and deploy replacement pods - Increase replicas (scale up pods)
# Enable your developers to Experiment - A/B testing recommender engines - k8s Canary deployments (requires app to support side-by-side 25%/75% split).
# Recreate with downtime - ?? # Rolling updates (no downtime) - Gradually version 1 to V1/V1.2 to version 2 (by monitoring
# Blue / Green Deployment - Scale down blue (but keep it around) - Green version 2.
# External Services - Switch between Dev and Prod effortlessly - Connect - Scala (Akka)
# Data Persistence - Statefulsets mongo-0, mongo-1, mongo-2 role mongo, mongo, mongo type leader, follower, follower type write, read, read
# Database Migrations - Create a flyaway (runs one time per db) Edson book from Redhat on how to write backward compatible database schema
# What's Next for Kubernetes - Kubevirt (virtual machine under mgmt) github/kubevirt - Day 2 Operations (coreos.com/operators) - Serverless - Knative github/knative - ServiceMesh - istio (service-to-service) traffic routing, resiliency/ - Kiali includes Jager tracing i.e. Traffic control, Service Resiliency, Chaos Monkey, ...
# Dark Launch - Route traffic to new version (while old version is still prod) # Secure by Default .... out of time https: http listing each istio service.