Skip to content

Latest commit

 

History

History
168 lines (130 loc) · 4.63 KB

README.adoc

File metadata and controls

168 lines (130 loc) · 4.63 KB
Raw

Quarkus Top 10 Kubernetes Features

This quarkus project is intended to follow the QuarkusTop 10 Kubernetes Features session that was delivered at j4k.io 2020 virtual conference.

Important
 To run this project on your own, follow the setup instructions in the yaml directory.

Top 10 features

Top 10 features shown include:

  1. One-step Kubernetes deployment

  2. Remote Live-coding in Kubernetes

  3. Direct ConfigMap and Secrets access

  4. Health endpoints

  5. Kuberentes Client API

  6. OpenMetrics support

  7. OpenTracing support

  8. One-step KNative deployment

  9. Functions as a Service with Funqy

  10. Efficiency

    • Fast startup time

    • Minimum memory utilization

Additional features beyond session

This project includes a few features/details that I couldn’t fit in the 45 minute slot, or even the 52 minutes I took to record it (whups, I went over!). These include:

  • Kubernetes Secrets

  • Database access to show Agroal datasource readiness check and Jaeger Tracing between the application and the database.

  • A custom liveness check, which makes the quarkus health-ui look a bit more populated.

Additional setup

In addition to the setup from the yaml directory, the j4k configmap and j4k-secret map must be configured:

Create j4k ConfigMap
kubectl create cm j4k \
    --from-literal=greeting="Hello ConfigMap"
Create j4k-secret secrets
kubectl create secret \
    generic \
    j4k-secret \
    --from-literal=db.user=quarkus \
    --from-literal=db.pass=quarkus

Deploy j4k application

Deploy the j4k application
cd j4k
mvn clean package -Dquarkus.kubernetes.deploy=true

Get and test endpoint

Test application endpoint
# We'll use this endpoint quite a bit, so let's save it
URL=`minikube service --url j4k`

# Get the persons (call database), should return JSON array of 3 persons
curl -i ${URL}/person

# Test the metrics endpoint
curl ${URL}/metrics

# Test the health endpoint
curl ${URL}/health

# Test Kube Client API (get pods)
curl ${URL}/kube/pods

# Stream pod events
curl ${URL}/kube/pods/stream

# 2nd terminal
kubectl scale --replicas 2 deployment/j4k

# Wait for pods to come up, then scale back down
kubectl scale --replicas 1 deployment/j4k

Open ${URL}/health-ui to test Health UI.

Open http://localhost:3000 to view grafana dashboard

Open http://localhost:16686 to view Jaeger UI

Knative and Funqy

Important
 This example uses quay.io, which creates a private image repository by default. If using Quay, go into the repository settings and make it a public repository. Using Docker Hub should work as-is.
Deploy Funqy
cd funqy
mvn \
    clean \
    package \
    -Dquarkus.kubernetes.deploy=true \
    -Dquarkus.kubernetes.deployment-target=knative

Once deployed, find the funqy service endpoint using the kn command line utility, and then curl the resulting endpoint. The URL shown will likely be differet than yours:

Invoke Funqy function
kn service list
curl -i http://funqy.default.10.98.14.70.nip.io/message

If you wait a minute or so, the number of pods will scale down to zero (using kubectl get pods -w). Scale up the pods dynamically by curling the endpoint again.

Startup time and memory consumption

Get the name of any running pod using kubectl get pods and run kubectl logs <POD> funqy. The startup time will show there. This example shows the function starting in 1.1 seconds using the JVM on loaded minikube instance:

Example log output
exec java -Dquarkus.http.host=0.0.0.0 -Djava.util.logging.manager=org.jboss.logmanager.LogManager -XX:+ExitOnOutOfMemoryError -cp . -jar /deployments/app.jar
__  ____  __  _____   ___  __ ____  ______
 --/ __ \/ / / / _ | / _ \/ //_/ / / / __/
 -/ /_/ / /_/ / __ |/ , _/ ,< / /_/ /\ \
--\___\_\____/_/ |_/_/|_/_/|_|\____/___/
2020-09-24 19:49:22,995 INFO  [io.quarkus] (main) funqy 1.0.0-SNAPSHOT on JVM (powered by Quarkus 1.8.1.Final) started in 1.132s. Listening on: http://0.0.0.0:8080
2020-09-24 19:49:22,997 INFO  [io.quarkus] (main) Profile prod activated.
2020-09-24 19:49:22,997 INFO  [io.quarkus] (main) Installed features: [cdi, funqy-http, kubernetes]

To view RSS (memory usage) of the JVM instance, run: .Get RSS value

ssh minikube
ps -ef | grep app.jar  # get the process id
ps -o rss,command <PID>
Example output
   RSS COMMAND
118160 java -Dquarkus.http.host=0.0.0.0 -Djava.util.logging.manager=org.jboss.logmana

The JVM process will use 118M of RAM.