Skip to main content
Version: 3.6.x

Access and Tools

Accessing logs​

There are many different kinds of logs that are accessible using the kubectl command line tool. The following sections show examples for accessing some of these logs.

Access pod logs​

To access the logs for a pod use the kubectl logs command.

  • For example, run the following command:

    POD=$(kubectl get pods -l 'app=ep-mysql' -o jsonpath='{.items[*].metadata.name}') \
    kubectl logs pod/${POD}
    note

    Ensure that you adjust the value ep-mysql to match the label of the pod.

Access cluster event logs​

Cluster event logs give you visibility into why nodes or pods are created or deleted.

  • Run the following command to view cluster events:

    kubectl get events

Access cluster logs​

  • To see logs for all pods and nodes in a cluster, run the following command:

    kubectl cluster-info dump

Accessing a shell on an EKS Cluster Member​

EC2 nodes that are members of CloudOps for Kubernetes EKS clusters can be accessed using AWS Session Manager. This capability is configured for nodes and node groups created with CloudOps for Kubernetes release 3.4.0 or higher. Specific permissions added to the NodeGroupPolicy AWS Identity and Access Management (IAM) policy enable this. For information about using Session Manager, see the AWS Session Manager documentation.

Accessing a shell inside a running container​

To be able to run shell commands inside a running container use the kubectl exec command.

  • For example, run the following command:

    POD=$(kubectl get -n ${NAMESPACE} pods -l 'app=ep-cm' -o jsonpath='{.items[*].metadata.name}') \
    kubectl exec -n ${NAMESPACE} -it ${POD} -- bash -c 'echo helloworld'
    note

    Ensure that you adjust the value app=ep-cm to match the label of the pod.

Creating secure tunnel to a port of a container​

To create a secure tunnel to a port on one of the containers use the kubectl port-forward command. For more information, see the Kubernetes documentation.

For example, the following command port-forwards a Cortex Pod running in the default namespace to port 5678 on your local machine:

POD=$(kubectl get pods \
        -l 'app=ep-cortex' \
        -o jsonpath='{.items[*].metadata.name}') \
kubectl port-forward pod/${POD} 5678:8080
tip

Using the kubectl port-forward command is the only way to access the OSGi Felix console on the Cortex application.

Checking Docker Pull Secrets for the cluster​

  • To check the Docker Pull Secrets for the cluster, run following command:

    kubectl get secret ep-registry-creds-secret -o json | \
    jq -r '.data.".dockerconfigjson"' | base64 -d

Listing configuration of components deployed through Helm​

  • To list the configuration of components deployed through Helm, run the following command: helm get manifest.

Deploy a bastion instance​

CloudOps for Kubernetes release 3.6.0 or higher includes the ability to deploy a bastion instance using the create-and-manage-bastion-instance Jenkins job. The bastion instance includes the eksctl, helm, and kubectl commands for interacting with your Kubernetes cluster, as well as SSH access to the cluster nodes.

To ensure the bastion instance has the necessary access in your environment, you will need to provide the following SSH keys:

  • A public SSH key for accessing the bastion instance
  • The private SSH key associated with the aws_eks_ssh_key variable defined in docker-compose.override.yml