Kubernetes Monitoring is a one-stop observability service that is designed for Kubernetes clusters. Kubernetes Monitoring provides IT developers and O&M engineers with a comprehensive observability solution that covers multiple aspects of Kubernetes clusters, such as metrics, traces, logs, and events.


  • Zero code intrusion: Kubernetes Monitoring uses bypass technologies to obtain abundant network performance data, without the need to instrument business code.
  • Language independence: Kubernetes Monitoring parses network protocols at the kernel level and supports all programming languages and frameworks.
  • High performance: Kubernetes Monitoring is developed based on the extended Berkeley Packet Filter (eBPF) technology. Kubernetes Monitoring can obtain abundant network performance data by consuming only a few resources.
  • Display of resource correlations: Kubernetes Monitoring displays correlations between resources in network topologies and resource topologies.
  • Data variety: Kubernetes Monitoring supports various forms of observable data, such as metrics, traces, logs, and events.
  • Integration: The Application Real-Time Monitoring Service (ARMS) console provides preset scenarios to integrate Kubernetes Monitoring with architecture discovery, Prometheus Service, internal network tests, alert configuration, and cluster management.


  • Resource performance monitoring

    Kubernetes Monitoring uses the eBPF technology to obtain the following performance data of containers without code intrusion: the number of requests, the number of failed requests, and the response time. Kubernetes Monitoring can efficiently discover performance issues within containers and pods. In addition, Kubernetes Monitoring can identify the services and controller workloads that are associated with the issues. The controller workloads include Deployments, StatefulSets, and DaemonSets. This improves troubleshooting efficiency. For more information, see View the overall monitoring information.

  • Cluster network topologies and network tracking
    • Kubernetes Monitoring can analyze network requests and automatically parse network protocols to generate network topologies. Kubernetes Monitoring supports the HTTP, Redis, Kafka, and MySQL protocols.
    • Each network topology displays the network performance between containers or between a container and an Alibaba Cloud service instance. You can efficiently identify performance issues within the relevant Alibaba Cloud services based on the topology.
    • By default, Kubernetes Monitoring stores the logs of failed requests and slow requests. Failed requests are the requests for which an HTTP status code greater than or equal to 400 is returned. Slow requests are the requests whose response time exceeds 500 ms. You can view the logs to troubleshoot the errors.

    For more information, see View the network topologies of a cluster.

  • Alert configuration

    Kubernetes Monitoring provides out-of-the-box alert templates. You can create alert rules based on preset alert templates or customize alert rules for specific Kubernetes clusters. When an alert rule is triggered, the system sends alert notifications to the specified contacts by using the specified notification methods. This reminds the contacts to take necessary measures to resolve the issue. For more information, see Create alert rules for Kubernetes Monitoring.

  • Monitoring of ECS resources and workload resources
    • Kubernetes Monitoring monitors the basic metrics of Elastic Compute Service (ECS) instances to ensure that the instances have sufficient resources. The basic metrics include CPU utilization, memory usage, and disk usage.
    • Kubernetes Monitoring monitors the basic metrics of pods and containers to ensure that the workload resources are sufficient. The basic metrics include the CPU utilization, CPU request rate, CPU limit rate, memory usage, memory request rate, memory limit rate, and disk usage.

    For more information, see Pod details and Container details.

  • Internal network tests

    You can use the health check feature to periodically test the connectivity of monitored services. This feature allows you to check the health status of the monitored services and handle detected exceptions at the earliest opportunity. For more information, see Perform inspections.

    The agent check feature allows you to use Blackbox Exporter to collect monitoring data that is related to protocols such as HTTP, DNS, TCP, and ICMP. For more information, see Protocol-related metrics.

  • Cluster management

    The cluster management feature allows you to set the protocol ports, metrics, and traces that are collected by Kubernetes Monitoring. You can restart an agent if necessary. For more information, see Cluster configuration and Agent management.


Compared with open source Kubernetes monitoring services, Alibaba Cloud Kubernetes Monitoring has the following advantages:

  • Innovative display of resource correlations and interactions

    Kubernetes Monitoring generates network topologies by monitoring network requests. In each topology, you can check the dependencies between services.

    2D container topology
  • Data diversity

    Kubernetes Monitoring displays various types of data in a visualized manner. In addition, Kubernetes Monitoring displays data from different dimensions. For example, you can view the number of errors and the information about traces with errors.

    Figure 1. Number of errors
    Number of errors
    Figure 2. Information about a trace with errors
    Information about a trace with errors
  • Centralized feature management

    The ARMS console provides preset scenarios to integrate Kubernetes Monitoring with architecture discovery, Prometheus Service, internal network tests, alert configuration, and cluster management.

    • You can store and query metrics by using Prometheus Service.
    • You can enable the health check feature for services.
    • You can enable the alert notification feature.
    • You can configure collection metrics.

Contact us

If you have any questions in the Kubernetes Monitoring, please contact the Kubernetes Monitoring Q&A DingTalk group (group number: 35568145) for help.