Oracle Cloud Infrastructure Documentation

Monitoring Clusters

Find out how to monitor the clusters, node pools, and nodes you've created using Kubernetes Engine (OKE).

Having created a cluster, you can monitor the overall status of the cluster itself, and the nodes and node pools within it.

In addition to monitoring the overall status of clusters, node pools, and nodes, you can monitor their health, capacity, and performance at a more granular level using metrics , alarms , and notifications. See Kubernetes Engine (OKE) Metrics.

Using the Console

To monitor a Kubernetes cluster:

  1. Open the navigation menu and click Developer Services. Under Containers & Artifacts, click Kubernetes Clusters (OKE).

  2. Choose a Compartment you have permission to work in.

    The Status column on the Cluster List page shows a summary status for each individual cluster and its control plane nodes. Clusters can have one of the following statuses:

    Cluster Status Explanation Possible Reason
    Creating Cluster is in the process of being created. Application is being deployed.
    Active Cluster is running normally. Control plane nodes are running normally.
    Failed Cluster is not running due to an unrecoverable error.

    Possible reasons:

    • a problem setting up load balancers
    • conflicts in networking ranges
    Deleting Cluster is in the process of being deleted. Application no longer required, so resources in the process of being released. Application no longer required, so resources in the process of being released.
    Deleted Cluster has been deleted. Application no longer required, so resources have been released. Application no longer required, so resources have been released.
    Updating Version of Kubernetes on the control plane nodes is in the process of being upgraded. A newly supported version of Kubernetes has become available.

    Note that the cluster's summary status is not necessarily directly related to the status of node pools and nodes within the cluster.

  3. On the Cluster List page, click the name of the cluster for which you want to see detailed status.

  4. Display the cluster's Metrics tab to see more granular information about the health, capacity, and performance of the cluster. See Kubernetes Engine (OKE) Metrics.
  5. Display the Node Pools tab to see the summary status of each node pool in the cluster.

  6. On the Node Pools tab, click the name of a node pool for which you want to see detailed status.

    Node pools statuses include the following:

    Node Pool Status Explanation Possible Reason
    Creating Node pool is in the process of being created. Cluster is in the process of being created.
    Active Node pool is running normally. Worker nodes in the node pool are running normally.
    Deleted Node pool has been deleted. Application no longer required, so resources have been released.
    Needs attention At least one of the nodes in the node pool has an issue that requires investigation. An attempt to terminate a node in the node pool (for example, to scale down the node pool) failed because the node could not be drained and cordoned within the eviction grace period. See Notes on Cordoning and Draining Managed Nodes Before Termination.
  7. Display the node pool's Metrics tab to see more granular information about the health, capacity, and performance of the node pool. See Kubernetes Engine (OKE) Metrics.

  8. Display the Nodes tab to see the summary status of each worker node in the node pool.

    Worker nodes can have one of the following statuses:

    Node Status Explanation Possible Reason
    Creating Node is being created. Compute instance in the process of being created.
    Active Node is running normally. Node is running normally.
    Updating Node is in the process of being updated.

    Kubernetes Engine is performing an operation on the node.

     
    Deleting Node is in the process of being deleted. Application no longer required, so resources in the process of being released.
    Deleted Node has been deleted. Application no longer required, so resources have been released.
    Inactive Node still exists, but is not running. Compute resource has a status of Stopped, Stopping, or Down For Maintenance.
  9. Click View Metrics beside a worker node to see more granular information about the health, capacity, and performance of that node. See Kubernetes Engine (OKE) Metrics.