Access Elasticsearch clusters using Cerebro

Note

Prior to enabling Cerebro, verify that your Container Cloud cluster has minimum 0.5-1 GB of free RAM and 1 vCPU available.

Cerebro is a web UI for Elasticsearch, allowing visual inspection of and interaction with an Elasticsearch cluster, useful for evaluating its health and convenient debugging. Cerebro is disabled by default. Mirantis recommends that you enable Cerebro if needed, for example, if your Elasticsearch cluster encounters an issue, and disable it afterward.

To enable or disable Cerebro, set the logging.cerebro parameter to true or false as described in Configure StackLight and StackLight configuration parameters.

To access the Cerebro web UI:

  1. Obtain the host IP address:

    1. Log in to the Container Cloud web UI with writer or operator permissions and switch to the required project.

    2. From the Clusters page, click the required cluster.

    3. Click the action icon in the last column of any machine of the manager type.

    4. Click Machine info and copy the Host IP.

  2. Log in to a local machine where your management cluster kubeconfig is located and where kubectl is installed.

  3. Obtain the cluster network CIDR:

    kubectl get cluster kaas-mgmt -o jsonpath='{.spec.clusterNetwork.services.cidrBlocks}'
    
  4. Create an SSH tunnel to the host, for example, using sshuttle:

    Note

    This step requires SSH access to Container Cloud hosts.

    sshuttle -r ubuntu@<HOST_IP> <CIDR>
    
  5. Obtain the Cerebro IP address:

    kubectl get svc -n stacklight cerebro -o jsonpath='{.spec.clusterIP}'
    
  6. Paste the Cerebro IP address in a web browser.