I do not maintain this chart and it is likely not to work. I suggest using https://github.com/confluentinc/cp-helm-charts.
This chart let you install Kafka manager on a Kubernetes cluster using the kafka-manager-docker image.
This section allow you to configure which Kafka cluster the manager will monitor. You can skip this step and configure this using the webUI, but you will need to configure a Zookeeper backend (See below).
kafka:
clusters:
- # You can embed any configuration you need, variables have the same
# name than the web ui "Add Cluster" form variables.
name: "default"
# MANDATORY VALUE
# Uri at which Kafka Zookeeper's can be contacted.
zkHosts: "kafka-zookeeper:2181"
kafkaVersion: "0.9.1"
Kafka manager need a Zookeeper
cluster to work, this chart let you 2 options
for it's configuration :
- Use the Kafka internal
Zookeeper
cluster (by default if you have configured your clusters) . - Use an external cluster (
values.yaml
entries below)
kafkaManager:
useKafkaZookeeper: false
zkHosts: "some-zookeeper:2181"
- Clone the repository
- Configure the application by tweaking
values.yaml
helm install .
Note that the helm installation/upgrade can be slow because the cluster configuration
must be done at runtime, we are thus using a helm post-install
hook to send
cluster add requests from a alpine-curl
image. Helm won't exit before the
post-install
job succeeded.
If you are running into an Error: timed out waiting for the condition
, this probably mean that the kafka-manager
deployment isn't healthy,
thus preventing the hook termination.
Once the chart is deployed, it will give you some informations on how to access your application:
Access to the application using the ingress :
http://kafka-manager.local
Access to the application using service :
export POD_NAME=$(kubectl get pods --namespace default -l "app=kafka-manager,release=torpid-yak" -o jsonpath="{.items[0].metadata.name}")
echo "Visit http://127.0.0.1:8080 to use your application"
kubectl port-forward $POD_NAME 8080:9000