gcp / gke / k8s microservices managing microservices with kubernetes 08.17.2016 Austin GCPUG Lou Scalpati CTO RxWiki
microservices What are Microservices generally REST APIs returning json organized around function: a user service, crud for db app or a url shortening service small units of an app but are standalone Why Microservices? modular easy to deploy scales independently Reduces maintenance issues uncouples components of your app - changes can be deployed without having to test entire codebase reusable usually a single developer or small team can create / manage mixed language environment
monolithic vs microservices https://www.safaribooksonline.com/library/view/kubernetes-cookbook/9781785880063/ch05s02.html
the 12 Factor App (or 12 factor microservice) I. Codebase (git) One codebase tracked in revision control, many deploys II. Dependencies (vendoring) Explicitly declare and isolate dependencies III. Config (docker) Store config in the environment IV. Backing services (docker/k8s) Treat backing services as attached resources V. Build, release, run (docker/k8s) Strictly separate build and run stages VI. Processes (docker/k8s) Execute the app as one or more stateless processes VII. Port binding (docker/k8s) Export services via port binding VIII. Concurrency (docker/k8s) Scale out via the process model IX. Disposability (docker) Maximize robustness with fast startup and graceful shutdown X. Dev/prod parity (docker/k8s) Keep development, staging, and production as similar as possible XI. Logs (docker/k8s) Treat logs as event streams XII. Admin processes (k8s) Run admin/management tasks as one-off processes
what is Kubernetes (k8s) What it does application configuration service Discovery managing Updates monitoring containers PAAS that manages your application deployment and operations Commonly used components pods deployments replicasets /replication controllers services labels (tag) driven
k8s Pods and Services Pods tightly coupled containers scaled together shared: Namespace Volumes IP short life span - spin up and down quickly Services defines a logical set of pods defines access policy to pods really a microservice Targets pods by label selector base unit for service discovery More permanent than pods
the architecture (the obligatory diagram) internet Service Load Balancer Service Service Pod Pod Pod Pod Pod Pod
deployments and ReplicaSets Deployments combines pods and scaling policies declarative updates for Pods and Replica Sets describe the desired state deployment controller will: change the actual state to the desired state Control rate of change Rollback to an earlier Deployment revision History ReplicaSets the next-generation Replication Controller works as part of a deployment can be used standalone (but why?)
creating a deployment apiversion: extensions/v1beta1 kind: Deployment metadata: name: nginx-deployment spec: replicas: 3 template: metadata: labels: app: nginx spec: containers: - name: nginx image: nginx:1.7.9 ports: - containerport: 80 create a deployment $ kubectl create -f docs/user-guide/nginx-deployment.yaml --record deployment "nginx-deployment" created get deployment status $ kubectl get deployments NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE nginx-deployment 3 3 3 3 18s get ReplicaSet Status $ kubectl get rs NAME DESIRED CURRENT AGE nginx-deployment-2035384211 3 3 18s
fun with deployments (updates and rollouts) updating a deployment (rollout) $ kubectl set image deployment/nginx-deployment nginx=nginx:1.9.1 deployment "nginx-deployment" image updated or edit the yaml (rollout) and $ kubectl edit deployment/nginx-deployment deployment "nginx-deployment" edited status of deployment (rollout) $ kubectl rollout status deployment/nginx-deployment deployment nginx-deployment successfully rolled out shows the status (new and old) $ kubectl get rs NAME DESIRED CURRENT AGE nginx-deployment-1564180365 3 3 6s nginx-deployment-2035384211 0 0 36s
deployment history check history $ kubectl rollout history deployment/nginx-deployment deployments "nginx-deployment": REVISION CHANGE-CAUSE 1 kubectl create -f docs/user-guide/nginx-deployment.yaml --record 2 kubectl set image deployment/nginx-deployment nginx=nginx:1.9.1 3 kubectl set image deployment/nginx-deployment nginx=nginx:1.91
deployment history (details of a point in time) $ kubectl rollout history deployment/nginx-deployment --revision=2 deployments "nginx-deployment" revision 2 Labels: app=nginx pod-template-hash=1159050644 Annotations: kubernetes.io/change-cause=kubectl set image deployment/nginx-deployment nginx=nginx:1.9.1 Containers: nginx: Image: nginx:1.9.1 Port: 80/TCP QoS Tier: cpu: BestEffort memory: BestEffort Environment Variables: <none> No volumes.
roll back (this is why to use deployments) undo! $ kubectl rollout undo deployment/nginx-deployment deployment "nginx-deployment" rolled back roll back to a point in time! $ kubectl rollout undo deployment/nginx-deployment --to-revision=2 deployment "nginx-deployment" rolled back Note: Kubectl rolling update updates Pods and Replication Controllers in a similar fashion. But Deployments are recommended, since they are declarative, server side, and have additional features, such as rolling back to any previous revision even after the rolling update is done.
load balancing options GCP network load balancer automatic defined in service connects directly to service layer 3 no SSL GLBC k8s L7 Load balancer Still beta Layer 7 configure ingress objects does not clean up all network resources no SSL (?? ingress object) GCP HTTP/HTTPS Load Balancer manual config SSL Certificate process?? nginx defacto standard reverse proxy supports SSL layer 7 needs configuration easy to deploy haproxy versatile L3,4,5,6,7 designed for network proxy what I use many others
resources the 12 factors: http://goo.gl/ay6ov K8s services: http://goo.gl/gcuj8j k8s deployments: http://goo.gl/dtgty1 k8s microservices course (free): https://goo.gl/tfkns3