Docs

Volcengine Kubernetes Engine

Full stack observation

Enabling full-stack observability

Full-stack observability provides a metric collection method without code modification by using eBPF, which achieves metric and trace monitoring for cluster networks, applications, and services. This topic describes how to configure the full-stack observability feature.

Prerequisites

Cloud-native observability has been enabled. For details, see Enabling observability.

TLS has been activated. To activate this service, log in to the TLS console and activate the service based on the on-screen prompts.

A log project has been created in TLS. For details, see Log projects

APMPlus has been activated. To activate this service, log in to the APMPlus console and activate the service based on the on-screen prompts.

Enabling observability

Step 1: Enable observability

In the left navigation bar, click Cluster. On the page that appears, find the target cluster and click its name.

In the left navigation bar of the cluster management page, click Observability settings and click the Full-stack observability tab.

In the Network and application observability section, click Enable, and configure network and application observability.

Prometheus workspace: This section displays the VMP workspace already associated with the cluster. For details, see Operation and maintenance configuration. Click the workspace name to change the workspace in the pop-up dialog box.

Warning

If you change the VMP workspace, the workspace associated with the cluster will change. Proceed with caution.

Enable status check for VKE observability: Check whether cloud-native observability is enabled in the cluster. If this feature is not enabled, you need to enable cloud-native observability first. For details, see Enabling observability.

Step 2: Configure the probe

Configure Microscope Agent probe: Click Configuration parameters to configure the Microscope Agent probe.

Specify the parameters of the probe, such as the specifications and the number of replicas.

Parameters	Description
Built-in parameters	The system provides multiple preset parameters to help you quickly specify the parameters of the probe, including Small instance type, Medium instance type, Large instance type, and Blank configuration. If you select Built-in parameters, the system automatically fills in the parameter configurations that correspond to the probe.
Microscope Agent	Configure the resource requirements of the Microscope Agent add-on, including the CPU request, CPU upper limit, memory request, and memory upper limit.
OTel Collector	Configure the resource requirements and autoscaling configurations for the OTel Collector add-on, including the CPU request, CPU upper limit, memory request, and memory upper limit. Resource: Configure the resource requirements of the OTel Collector add-on, including the CPU request, CPU upper limit, memory request, and memory upper limit. Autoscaler: Configure the autoscaling settings for the OTel Collector add-on, including the minimum number of instances, maximum number of instances, CPU utilization, and memory utilization.
Microscope Operator	Configure the resource requirements of the Microscope Operator add-on, including the CPU request, CPU upper limit, memory request, memory upper limit, and the initial number of instances.

After you specify the parameters of the probe, click Next: Check before deploy. The system will automatically check whether the probe can be installed with the current cluster resource utilization. When cluster resources are insufficient, you need to first scale out the cluster.

Click Submit to complete the probe configuration.

Step 3: Configure the observability scope

In the workload list, select the workloads for which you want to enable network and application observability.

Click Select label to filter workloads by label or annotation.

tip

You can use multiple labels or annotations to filter workloads. Multiple labels and annotations are in an AND relationship.

You can directly click and select the corresponding label or annotation in the Label column or Annotation column of the workload list.

After configuration is complete, newly deployed workloads in the cluster will not be automatically added to the observability scope. You can also perform manual configuration.

After all configurations are complete, click Enable to enable network and application observability and install the managed probe.

Configuring collection

Configuring metric collection

After network and application observability is enabled, you can configure metric collection and select the network and application metrics that you need to collect.

In the left navigation bar of the cluster management page, click Observability settings and click the Full-stack observability tab.

Find the Network and application observability section and click Edit settings. Click the Metric tab, and click Edit to configure eBPF metrics.

3. After configuration is complete, click Confirm to complete the metric configuration.

Configuring trace data collection

In the left navigation bar of the cluster management page, click Observability settings.

Find the Network and application observability section and click Edit settings. Click the Chain tab. You can collect data traces by using APMPlus or TLS.

Parameters	Description
Storage settings	Configure the storage service used after trace data collection. APMPlus or TLS is supported.
Dependency check	The system will automatically check whether the data collection storage service (APMPlus or TLS) is authorized for use and enabled. If the service fails to pass the check, check whether you have activated the storage service.
Configuring collection
Project	If you set the Storage settings parameter to TLS, select a project from the drop-down list. Project management is a resource management method. For details, see Project overview.
Log project	If you set the Storage settings parameter to TLS, select a log project from the drop-down list.
Application protocol	Select the application protocol to enable trace collection. HTTP, HTTPS, DNS, gRPC, and Redis are supported.
Slow request threshold	Configure the slow request threshold for different protocols.

After configuration is complete, click Confirm to complete the trace data collection configuration.

Performing operation and maintenance on probes

After installing the probe, you can monitor, maintain, and restart it.

Monitoring probes

In the left navigation bar of the cluster management page, click Observability settings.

Find the Network and application observability section and click Edit settings. Click the Probe tab to view the running status and resource usage of the Microscope-agent and Otel Collector probes.

Restarting probes

Click Restart in the Actions column to restart the specified probe.

Viewing metrics and topologies

To view network and performance metrics, see Resource retrieval.

To view resource topologies, see Resource topologies.

Last updated: 2026.07.24 10:53:30