This directory contains a number of example that demonstrate different features of Kroxylicious in a Kubernetes environment.
The instructions for running the examples are given below.
| example | showcases |
|---|---|
| record-encryption | kroxylicious providing encryption-at-rest. |
| multi-tenant | kroxylicious providing multi-tenancy - present one kafka cluster as if it were many. |
| portperbroker_plain | kroxylicious+strimzi using plain connections upstream/downstream. |
| snirouting_tls | kroxylicious+strimzi using TLS upstream/downstream connections with SNI based routing. |
If you want build your own kroxylicious images you'll additionally need:
- Docker engine or podman
- Access to a container registry such as quay.io or docker.io with a public accessible repository within the registry named
kroxylicious.
Kroxylicious can be containerised and run on Minikube against a Strimzi managed Kafka cluster.
To run using pre-built Kroxylcious images:
./scripts/run-example.sh ${kubernetes_example_directory}where ${kubernetes_example_directory} is replaced by a path to an example directory e.g. ./kubernetes-examples/portperbroker_plain.
To use an alternative image for Kroxylicious, set the KROXYLICIOUS_IMAGE environment variable.
KROXYLICIOUS_IMAGE=quay.io/kroxylicious/kroxylicious:x.y.z ./scripts/run-example.sh ${kubernetes_example_directory}This run-example.sh script does the following:
- starts minikube (if necessary)
- installs cert manager, vault and strimzi as necessary.
- installs a Kafka cluster using Strimzi into minikube
- installs kroxylicious into minikube, configured to proxy the cluster
All the examples configure Kroxylicious to emit metrics on the default port 9190, however
there is nothing configured to scrape them.
You can make an adhoc observation of the metric like this:
kubectl run --restart=Never --attach=true --rm=true -n kroxylicious metrics --image=busybox -- wget -O - http://$(kubectl get pods -n kroxylicious -l app=kroxylicious -o jsonpath="{.items[0].status.podIP}"):9190/metricsFor anything above an adhoc observation, use Prometheus to consume the metrics and something like Grafana to visualise them.