Skip to content

Service Details

In Performance Metrics or Unified Catalog > Service, click a service to enter its details page.

When entering from cross-workspace query results, the details use the source workspace of the selected service and display the currently viewed workspace and service name. Services with the same name in different workspaces are queried separately; in cross-workspace read-only mode, you cannot modify service configuration, create monitors, or bind views. For related conditions, see Cross-Workspace Query Notes.

Components

The service details page consists of tabs such as Summary and Resource Invocation, as well as elements such as the Time Widget.

Summary

Displays the number of error spans, error requests, error logs, and error trends filtered by Service, Environment, and Version. It also supports analysis views from different dimensions:

  • Performance Analysis: covers key metrics including average response time, service request response time, slow call count, and service request distribution;
  • Resource Analysis: displays the TOP 50 data list of resource invocation analysis;
  • Anomaly Analysis: presents anomalies proactively detected by the system and their related context, including the TOP 10 distribution of monitors associated with events and the distribution of associated log statuses.

For more details, see Analysis Dashboard.

Resource Invocation

Displays resource invocation analysis associated with the service, filtered by the Environment, Version, and K8s Cluster dimensions. The left side shows the resource invocation Top List under the filter conditions; click any resource entry, and the right side displays the detailed view of that resource.

On the left side, the resource invocation Top List associated with the current service is displayed. You can switch to view the invocation Top List of all resources.

  • Entry Resource: The Root Span in a Trace, representing the initial request entering the system from outside, and the direct service entry provided by the system to the outside. A Trace has only one Root Span;

  • All Resources: Displays all invocation nodes in the trace, including the entry and all its downstream dependencies. The data volume is large. A Trace contains multiple Spans.

Click a resource entry, and the right side displays:

  • Time Series Trend Chart: displays core metrics in chart form, including invocation count, response time, error rate, response time distribution, and request count by status status: ok,error;
  • Complete Trace List: lists trace data related to this resource; click to expand the details of a single trace;

  • Resource Upstream/Downstream Topology: centered on the current resource, displays nodes and connections that have direct or indirect invocation relationships with it, and provides the following interaction capabilities:

    • Fill topology node colors based on the three dimensions of event status, P99 response time, and request error rate;
    • Hover over a node to directly display key metric data;

    • Click any resource node in the topology to directly view its upstream/downstream invocations, related logs, user access, and related events, or create a Monitor based on this node;

    • By default, multiple levels of upstream/downstream invocation relationships are displayed, and nodes that can be further expanded are indicated;

    • Use the View Upstream/Downstream button to continue expanding and analyze, level by level, the nodes and connections that have direct or indirect invocation relationships with the current resource;
    • Resources at the same level are sorted by response time, making it easier to locate high-latency resources;
    • Click an invocation connection to view the key metrics and trends of that invocation relationship and analyze upstream/downstream invocation performance;
    • After adjusting or expanding the topology, click Restore to reset the node layout and display scope to the initial state;
    • Click Full Screen to enlarge or restore the topology view in one click.

Traces

The system automatically filters trace data associated with the current service. You can perform all operations available in the Trace Explorer on this page.

For more details, see Traces.

Error Tracking

The system automatically filters error data associated with the current service. You can perform all operations available in the Error Tracking Explorer on this page.

For more details, see Error Tracking.

Service Map

By default, the upstream and downstream invocation relationships of the current service are displayed as rectangular metric cards. Each card shows the service name, request rate, P99 response time, and error rate, with the current service highlighted. Cross-workspace services are also marked with their source workspace; services with the same name in different workspaces are displayed separately.

Invocation connections in the upstream/downstream topology display request count, error count, and P99 response time. Node metrics describe the service itself, while connection metrics describe the corresponding invocation relationship, making it easy to distinguish service performance from inter-service invocation. Missing metrics are displayed as -, and actual zero values are displayed as 0.

After searching for another service in the current Service Map, the system focuses on the invocation relationship between the current service and the target service, hiding nodes and connections unrelated to that relationship, so you can quickly analyze the dependency path between specified services.

You can also switch to the full-path or infrastructure service invocation topology. In the Full Path view, you can view service invocation relationships through the grouping feature, making it easier to analyze dependencies between services by group.

The upstream/downstream view uses fixed-size cards and does not provide a setting for calculating node size by request count; the full-path view retains the original node size and grouping rules.

The following figure shows an example of the upstream/downstream view:

Rectangular metric cards and invocation relationships in the service upstream/downstream topology

For more details, see Service Map.

Infrastructure Honeycomb

In the Service Map, switch to Infrastructure to select Host or Pod and view resources associated with the current service. Each resource is displayed as an equal-sized honeycomb cell. The current service name remains in the title bar; it does not occupy a honeycomb cell, and no connections are drawn between the service and resources.

  • Use Fill by Metric to select a metric and color the honeycomb cells according to the corresponding metric value;
  • Hover over a honeycomb cell to view the resource name and metrics; click to open the details of the corresponding host or Pod;
  • Use the mouse wheel to zoom and drag the canvas; click Full View to restore the initial view;
  • Infrastructure mode does not provide Service Map configurations such as node size, association highlight range, or grouping. After switching back to Upstream/Downstream or Full Path, the corresponding setting entries are restored.

The honeycomb view only adjusts the display form of resources; associated resources, metric statistics, and query scope follow the original rules. The standalone Infrastructure tab on the page is still used to view the metrics and lists described below.

Infrastructure honeycomb view of hosts associated with a service

Infrastructure

Based on the current service, displays key metric data of associated infrastructure dependencies:

  • Host: CPU usage, system load, process count, disk I/O time, disk IOPS, etc.;

  • Pod: CPU usage, memory usage, traffic, etc.

The corresponding list is displayed below. Click an item to expand the slide-out details page to view complete information.

The system automatically filters log, Profile, and event data associated with the current service.

Displays custom views bound to the current service.