Skip to content

README.md

Latest commit

 

History

History
356 lines (243 loc) · 12.6 KB

README.md

File metadata and controls

356 lines (243 loc) · 12.6 KB
Raw

Run Kpow Hourly for Apache Kafka on AWS Marketplace with Kubernetes

Kpow is the all-in-one toolkit to manage, monitor, and learn about your Kafka resources.

This Helm chart is designed for the Kpow Hourly listing on AWS Marketplace. It uses a custom AWS Marketplace container that integrates with AWS to verify subscriptions and meter your usage. The container is automatically licensed to the subscribing AWS account, which is billed directly for the subscription. There is no need to arrange a separate license with us if you subscribe to a Kpow product through the AWS Marketplace.

Helm Charts

This Helm chart is for the Kpow Hourly offering on AWS Marketplace.

Prerequisites

The minimum information Kpow requires to operate is:

  • License Details: No license required—billing is handled automatically through your AWS account.
  • Kafka Bootstrap URL

See the Kpow Documentation for a full list of configuration options.

Kubernetes

Create a Service Account with IAM permissions

eksctl create iamserviceaccount \
    --name kpow \
    --namespace factorhouse \
    --cluster <ENTER_YOUR_CLUSTER_NAME_HERE> \
    --attach-policy-arn arn:aws:iam::aws:policy/AWSMarketplaceMeteringRegisterUsage \
    --approve \
    --override-existing-serviceaccounts

You can now deploy Kpow to EKS using this Service Account, which includes an IAM Role with the AWSMarketplaceMeteringRegisterUsage policy attached.

Configure Kubernetes/EKS

You need to connect to a Kubernetes environment before you can install Kpow.

The following examples demonstrate installing Kpow in Amazon EKS.

aws eks --region <your-aws-region> update-kubeconfig --name <your-eks-cluster-name>

Updated context arn:aws:eks:<your-aws-region>:123123123:cluster/<your-eks-cluster-name> in /your/.kube/config

Confirm Kubernetes Cluster Availability

kubectl get nodes

NAME                              STATUS   ROLES    AGE     VERSION
ip-192-168-...-21.ec2.internal   Ready    <none>   2m15s    v1.32.9-eks-113cf36
...

Run Kpow in Kubernetes

Configure the Kpow Helm Repository

Add the Factor House Helm Repository in order to use the Kpow Helm Chart.

helm repo add factorhouse https://charts.factorhouse.io

Update Helm repositories to ensure you install the latest version of Kpow.

helm repo update

Start a Kpow Instance

Start Kpow with config from '--set env.XYZ'

When using helm install, you can pass configuration with the --set env.XYZ flag. This requires careful formatting for certain values.

Some fields, particularly integers and strings containing quotation marks, require quoting. You may also need to escape special characters (like commas or nested quotes) with a backslash (\). For more details, see Helm's documentation on The Format and Limitations of --set.

The following example shows how to install Kpow from the command line, highlighting how to handle escaped commas and quotes:

helm install kpow factorhouse/kpow-aws-hourly \
  --set serviceAccount.create=false \
  --set serviceAccount.name=kpow \
  --set env.BOOTSTRAP="b-1.<cluster-name>.<cluster-identifier>.c8.kafka.us-east-1.amazonaws.com:9096" \
  --set env.SECURITY_PROTOCOL="SASL_PLAINTEXT" \
  --set env.SASL_MECHANISM="PLAIN" \
  --set env.SASL_JAAS_CONFIG="org.apache.kafka.common.security.plain.PlainLoginModule required username=\"user\" password=\"secret\";" \ # <-- note the escaped quotes
  --create-namespace --namespace factorhouse

NAME: kpow
LAST DEPLOYED: Mon May 31 17:22:21 2021
NAMESPACE: factorhouse
STATUS: deployed
REVISION: 1
NOTES:
1. Get the application URL by running these commands:
  export POD_NAME=$(kubectl get pods --namespace factorhouse -l "app.kubernetes.io/name=kpow,app.kubernetes.io/instance=kpow" -o jsonpath="{.items[0].metadata.name}")
  echo "Visit http://127.0.0.1:3000 to use your application"
  kubectl --namespace factorhouse port-forward $POD_NAME 3000:3000

Start Kpow with Environment Variables from a ConfigMap

You can configure Kpow with a ConfigMap of environment variables as follows:

helm install kpow factorhouse/kpow-aws-hourly \
  --set envFromConfigMap=kpow-config \
  --create-namespace --namespace factorhouse

This approach requires a ConfigMap named kpow-config to already exist in the factorhouse namespace. To configure Kpow with a local ConfigMap template, see Configuring with an Existing ConfigMap. You can also refer to kpow-config.yaml.example for a sample ConfigMap manifest.

For general guidance, see the Kubernetes documentation on configuring all key-value pairs in a ConfigMap as environment variables.

Manage a Kpow Instance

Set the $POD_NAME variable and test the Kpow UI

Follow the notes instructions to set the $POD_NAME variable and configure port forwarding to the Kpow UI.

export POD_NAME=$(kubectl get pods --namespace factorhouse -l "app.kubernetes.io/name=kpow,app.kubernetes.io/instance=kpow" -o jsonpath="{.items[0].metadata.name}")
echo "Visit http://127.0.0.1:3000 to use your application"
kubectl --namespace factorhouse port-forward $POD_NAME 3000:3000

Kpow is now available on http://127.0.0.1:3000.

Check the Kpow Pod

kubectl describe pods --namespace factorhouse

Name:         kpow-9988df6b6-vvf8z
Namespace:    factorhouse
Priority:     0
Node:         ip-172-31-33-42.ap-southeast-2.compute.internal/172.31.33.42
Start Time:   Mon, 31 May 2021 17:22:22 +1000
Labels:       app.kubernetes.io/instance=kpow
              app.kubernetes.io/name=kpow
              pod-template-hash=9988df6b6
Annotations:  kubernetes.io/psp: eks.privileged
Status:       Running

View the Kpow Pod Logs

kubectl logs $POD_NAME --namespace factorhouse

11:36:49.111 INFO  [main] kpow.system ? start Kpow
...

Remove Kpow

helm delete kpow --namespace factorhouse

Start Kpow with Local Changes

You can run Kpow with local edits to chart files to provide custom configuration.

Pull and Untar the Kpow Charts

helm pull factorhouse/kpow-aws-hourly --untar --untardir .

Make Local Edits

Make any edits required to kpow-aws-hourly/Chart.yaml or kpow-aws-hourly/values.yaml (adding volume mounts, etc).

Run Local Charts

The command to run local charts is slightly different, see ./kpow-aws-hourly rather than factorhouse/kpow-aws-hourly.

helm install kpow ./kpow-aws-hourly \
  <.. --set configuration, etc ..> \
  --create-namespace -namespace factorhouse

Configuring with an Existing ConfigMap

This is the recommended method for managing configuration separately from the Helm chart.

1. Prepare Your ConfigMap Manifest

Copy the example file (kpow-config.yaml.example), then edit it to set your desired metadata.name (e.g., kpow-config) and fill in your configuration under the data section.

cp ./kpow-aws-hourly/kpow-config.yaml.example kpow-config.yaml
# now edit kpow-config.yaml

2. Create the ConfigMap in Kubernetes

Before installing, use kubectl to create the ConfigMap object in your cluster from the file you just prepared.

kubectl apply -f kpow-config.yaml --namespace factorhouse

3. Install the Chart

Install the Helm chart, using --set to reference the name of the ConfigMap you just created. The --create-namespace flag will ensure the target namespace exists.

helm install kpow ./kpow-aws-hourly \
  --set envFromConfigMap=kpow-config \
  --create-namespace --namespace factorhouse

The Kpow pod will now start using the environment variables from your externally managed ConfigMap.

See kpow-config.yaml.example for an example ConfigMap file.

See the Kubernetes documentation on configuring all key-value pairs in a config map as container environment variables for more information.

Manage Sensitive Environment Variables

This helm chart accepts the name of a secret containing sensitive parameters, e.g.

apiVersion: v1
kind: Secret
metadata:
  name: kpow-secrets
data:
  SASL_JAAS_CONFIG: a3JnLmFwYWNoXS5rYWZrYS5jb21tb24uc2VjdXJpdHkucGxhaW4uUGxhaW5Mb2dpbk2vZHVsZSByZXF1aXJiZCB1c2VybmFtZT0iTFQ1V0ZaV1BRWUpHNzRJQyIgcGFzc3dvcmQ9IjlYUFVYS3BLYUQxYzVJdXVNRjRPKzZ2NxJ0a1E4aS9yWUp6YlppdlgvZnNiTG51eGY4SnlFT1dUeXMvTnJ1bTAiBwo=
  CONFLUENT_API_SECRET: NFJSejlReFNTTXlTcGhXdjNLMHNYY1F6UGNURmdadlNYT0ZXSXViWFJySmx2N3A2WStSenROQnVpYThvNG1NSRo=
kubectl apply -f ./kpow-secrets.yaml --namespace factorhouse

Then run the helm chart (this can be used in conjunction with envFromConfigMap)

See the Kubernetes documentation on configuring all key value pairs in a secret as environment variables for more information.

helm install kpow ./kpow-aws-hourly/ \
  --set envFromSecret=kpow-secrets \
  --set envFromConfigMap=kpow-config \
  --create-namespace --namespace factorhouse

Provide Files to the Kpow Pod

There are occasions where you must provide files to the Kpow Pod in order for Kpow to run correctly, such files include:

  • RBAC configuration
  • SSL Keystores
  • SSL Truststores

How you provide these files is down to user preference, we are not able to provide any support or instruction in this regard.

You may find the Kubernetes documentation on injecting data into applications useful.

Kpow Memory and CPU Requirements

The chart runs Kpow with Guaranteed QoS, having resource request and limit set to these values by default:

resources:
  limits:
    cpu: 2
    memory: 8Gi
  requests:
    cpu: 2
    memory: 8Gi

These default resource settings are conservative, suited to a deployment of Kpow that manages multiple Kafka clusters and associated resources.

When running Kpow with a single Kafka cluster you can experiment with reducing those resources as far as our suggested minimum:

Minimum Resource Requirements

resources:
  limits:
    cpu: 1
    memory: 2Gi
  requests:
    cpu: 1
    memory: 2Gi

Adjust these values from the command line like so:

helm install kpow ./kpow-aws-hourly/ \
     --set resources.limits.cpu=1 \
     --set resources.limits.memory=2Gi \
     --set resources.requests.cpu=1 \
     --set resources.requests.memory=2Gi \
     --create-namespace --namespace factorhouse

We recommend always having limits and requests set to the same value, as this set Kpow in Guaranteed QoS and provides a much more reliable operation.

Snappy compression in read-only filesystem

We preset an attribute for Snappy compression in read-only filesystems. It is disabled by default and can be enabled - modify the volume configuration if necessary.

ephemeralTmp:
  enabled: true
  volume:
    emptyDir:
      medium: Memory # Optional: for better performance
      sizeLimit: "100Mi" # Configurable size

Get Help!

If you have any issues or errors, please contact support@factorhouse.io.

Licensing and Modifications

This repository is Apache 2.0 licensed, you are welcome to clone and modify as required.