This example is for use with AWK EKS directly using a deployment file in place of helm.
In this example we are deploying as a distributed system. We are not deploying grafana, or any test apps.
Edit the 00-config.yml file to add the AWS Access and Secret keys. For details about how to set up the IAM role see our page in the docs AWS Permissions.
Configure kubectl to use the EKS cluster.
aws eks --region <AWS-Region> update-kubeconfig --name <Cluster-Name> --profile <AWS-CLI-Profile-Name>
Now deploy the pod to EKS.
kubectl apply -f .
Note: As this is deploying to AWS there could be increased costs at AWS.
To access deep when running in this mode you need to get the address from eks.
kubectl get service --namespace deep
This will output the service config including the external ip.
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
distributor LoadBalancer 10.100.16.87 a161316d4197043edbea4fd1b8ae86d0-1704341456.eu-west-1.elb.amazonaws.com 43315:31152/TCP 45s
ingester ClusterIP None <none> 43315/TCP,7946/TCP 44s
query-frontend LoadBalancer 10.100.33.208 a6fd86a8516e14548a8a786de24b4332-1338461574.eu-west-1.elb.amazonaws.com 3300:31750/TCP 45s
query-frontend-headless ClusterIP None <none> 3300/TCP,43315/TCP 44s
You can now use the value in 'EXTERNAL-IP' for the service 'query-frontend' to connect Grafana. Ensure you include the
protocol (http) and port (3300) http://a6fd86a8516e14548a8a786de24b4332-1338461574.eu-west-1.elb.amazonaws.com:3300
.
To connect a client to the service, you need to use the 'EXTERNAL-IP' for the service 'distributor'. Then set this to the value of 'DEEP_SERVICE_URL' for the client. (see the client docs for more info)
To remove the example run:
kubectl delete namespace deep
Note: This will only clean up the EKS cluster you will have to delete the S3 data yourself.