This topic describes the key features in PATROL for Kubernetes 1.0.00.
Dynamic discovery of Kubernetes Cluster entities
After you configure the Kubernetes cluster for monitoring, the KM automatically discovers the components, namespaces, nodes, controllers, and workloads of the cluster. The KM extracts the entire hierarchy of the cluster and displays it in the TrueSight console, indenting the various levels in the hierarchy.
It enables you to manage the cluster by monitoring the total number of nodes and running pods versus the total pods in the cluster. It also enables you to monitor the performance of the cluster by aggregating the CPU and memory resource utilization of the pods against the cluster capacity.
Monitoring of the cluster components
In component monitoring, the KM enables you to monitor the status (healthy, unhealthy, or unknown) of each component. It KM creates a container and lists all the components of the cluster below this container.
Monitoring of the cluster nodes
In node monitoring, the KM monitors the state of the node (ready) and various pressure conditions like disk pressure, memory pressure, and so on. It enables you to manage the node by monitoring the number of running pods versus the total pods. It discovers and monitors the newly added nodes dynamically and also enables you to monitor the CPU, Memory, Network, and FileSystem resources that are utilized by the node.
Monitoring of the cluster controllers
In controllers monitoring, the KM monitors the deployment, ReplicaSet, DaemonSet, and StatefulSet entities of the cluster. It enables you to monitor the overall performance of these controllers by aggregating the CPU and memory resource utilization of the pods against the cluster capacity. It also monitors the number of ready replicas versus the total replicas of the controller.
Monitoring of the cluster workloads
In workload monitoring, the KM monitors the availability of the pods and containers. It enables you to monitor the readiness of the pods (and containers) along with the state. It helps you in identifying the faulty applications by monitoring the number of container restarts per polling interval. It also monitors the CPU, memory, network, and storage metrics of the pods.
Filtering of cluster namespaces
Filtering allows you to include or exclude the namespaces from monitoring based on the requirements.
Device mapping of monitored entities
In device mapping, you can configure the KM to create a device for each Cluster, Node, and Pod in the TrueSight console. Devices are based on their hostnames and IP addresses.
Detailed annotation messages
The KM provides detailed annotation messages along with possible corrective actions in case of any failure. These annotations help you resolve the issue and configure the policy successfully.
The web-based configuration interface lets you manage the monitoring and filtering configuration easily. You can connect to the Kubernetes Cluster and start monitoring your environment simply by entering the Cluster name and providing either certificate-based or service-token based authentication details.