Observability

    The Observability addon provides system-level monitoring for KubeVela core and business-level monitoring for applications based on metrics, logging, and tracing data.

    The following describes observability capabilities in detail, and how to enable the observability addon and view various monitoring data.

    KubeVela observable capabilities are demonstrated through Grafana and provide system-level and application-level data monitoring.

    • KubeVela Core resource usage monitoring

    1) CPU, memory, and other usage and utilization data

    2) Graphical representation of CPU and memory usage and utilization over time (e.g. last three hours), and network bandwidth per second

    Observability - 图2

    Built-in metrics category II: KubeVela Core log monitoring

    1) Log statistics

    The observable page displays the total number of KubeVela Core logs, as well as the number of occurrences, frequency, overview of all logs that occur, and details by default.

    Observability - 图4

    2) Logging filter

    You can also filter the logs by filling keywords at the top.

    The observability plugin is installed with the vela addon command. Because this plugin relies on Prometheus, and Prometheus relies on StorageClass, the StorageClass varies to various Kubernetes distribution, so there are some differences in the installation command across Kubernetes distributions.

    Execute the following command to install the observability plugin. The steps are the same for similar clusters, like KinD.

    Kubernetes clusters provided by cloud providers

    Alibaba Cloud ACK

    The meaning of each parameter is as follows.

    • alertmanager-pvc-class

    The type of pvc required by the Prometheus alert manager, which is the StorageClass. On Alibaba Cloud, pick one from the StorageClass list.

    We set the value as .

    • alertmanager-pvc-size

    The size of the pvc needed by the Prometheus alert manager, on Alibaba Cloud, the minimum PV is 20GB, here it takes the value 20Gi.

    • server-pvc-size

    The size of the pvc required by the Prometheus server, same as alertmanager-pvc-size.

    • grafana-domain

    The domain name of Grafana, you can use either your custom domain name, or the cluster-level wildcard domain provided by ACK, . You can set the value as grafana.c276f4dac730c47b8b8988905e3c68fcf.cn-hongkong.alicontainer.com.

    Kubernetes clusters offered by other cloud providers

    Please change the following parameters according to the name and size specifications of the PVCs provided by different cloud provider’s Kubernetes clusters, and the domain rules.

    • alertmanager-pvc-class
    • alertmanager-pvc-size
    • server-pvc-class
    • server-pvc-size
    • grafana-domain

    Using username admin and the password above to login to the monitoring dashboard below.

    Get the monitoring url

    Visit http://127.0.0.1/dashboards and click on the corresponding Dashboard to view the various monitoring data introduced earlier.

    Observability - 图6

    • Kubernetes clusters provided by cloud providers

    Access the Grafana domain set up above directly to view the various monitoring data described earlier.

    On the Grafana home page, click on the console as shown to access the monitoring data for the appropriate category.

    The KubeVela Core System Monitoring Dashboard is the KubeVela Core system-level monitoring console. The KubeVela Core Logging Dashboard is the KubeVela Core logging monitoring console.