or Private accessible For short lived workloads like Bursts Scheduled work From € 1.05 for 1 CPU with 1 GB for 24 hours To € 5.22 for 4 CPU with 14 GB for 24 hours
Backend process & Front end With public available UI Run Continuously Over SSL with authentication Regularly new releases No downtime Auto scaling v2: v1:
or Private accessible 1 container instance only No High Availability No zero-downtime deployment No scale out Limited scale up No autoscaling No out of the box SSL support No cache for pulled containers Pay extra for Windows containers For short lived workloads like Bursts Scheduled work
Azure Red Hat OpenShift Azure Kubernetes Service (AKS) Azure Container Registry (ACR) Azure Container Instances (ACI) Run containers on Azure Azure Web App for Containers
exists Scale up Scale out Auto scaling Zero-downtime deployment SSL by default Authentication Identity Custom domains Hostingplan From € 3.59 for 1 CPU with 3.5 GB for 24 hours To € 14.37 for 4 CPU with 14 GB for 24 hours
all traffic Lots of containers Better density of our resources Make use of some CNCF projects - Mesh: Istio - Logging: Prometheus, Jaeger - Service discovery: CoreDNS - Messaging: NATS v3: v2: Backend process & Front end With public available UI Run Continuously Over SSL with authentication Regularly new releases No downtime Auto scaling
exists Scale up Scale out Auto scaling Zero downtime deployment SSL by default Authentication Identity Custom domains Scaling out is slow No optimal use of resources No firewall possibility (yet) Limited logging possibilities No health check functionality Only support for port 80 & 443 Not suitable for lots of containers Not portable Cannot debug
& scaling out Autoscaling Zero downtime deployment High Availability Public & Private endpoints Health management Enormous ecosystem Portable SSL Support* Identity management* Keyvault integration*
(Loadbalancer) Ingress controller IP-Address Ingress OurExternalAPI.com Ingress Myproject.com Ingress AdminSite.com Ingress Myproject.com/apis Works on ingress Auto request certificate Auto renewal https://github.com/jetstack/cert-manager DNS Zone
During rolling update deployment: Stop deployment During container startup No traffic Default endpoint Health endpoint (/health) Health endpoints returns != 200?
expandable standards but this still needs configuration. For example: • Deployments • Network Policies • Role Based Access Controls • Pod Security Policies • Pod Priority and more... Common integration points can be different across cloud providers • Authentication • Logging • Metrics • Storage
& scaling out Autoscaling Zero downtime deployment High Availability Public & Private endpoints Health management Enormous ecosystem Portable SSL Support* Identity management* Keyvault integration* Authentication A lot of management for a couple of containers Steep learning curve Not all Azure functionality is mature (yet) • Scale sets • Network policies • Multiple Node pools No turnkey configuration
scale in seconds High availability Per second billing Not Portable Container only (Preview v2 soon) Responsibility You Azure Application Deployment Hardware OS Patching Runtime upgrades Micro-billing Capacity planning Network & Storage