Enable APM Tracing¶

APM (Application Performance Monitoring) Tracing is used to trace the complete call path of requests in distributed systems.

It helps you:

Analyze performance bottlenecks: Locate the time consumption of each link (services, components) in the call chain.
Diagnose errors and exceptions: Quickly discover errors in the chain and their root causes.
Understand service dependencies: Visually display the call relationships between services.
Correlate monitoring data: Analyze comprehensively by linking with logs, metrics, and other data.

Integration Methods¶

Guance APM supports the OpenTracing protocol. You can choose one of the following methods to integrate.

DataKit is the all-in-one data collector of Guance, supporting multiple APM protocols.

Install DataKit.
Configure the APM collector.
- Enable and configure the corresponding DataKit collector based on the protocol used by your application (e.g., DDTrace, Jaeger, OpenTelemetry, SkyWalking, Zipkin).
Start/Restart DataKit: DataKit will automatically collect tracing data and report it to Guance.

You can integrate the APM SDK provided by Guance into your application, supporting Java, Python, Go, Node.js, .NET, etc.

For integration methods of each language SDK, refer to APM SDK.

After successful integration, you can view and analyze APM tracing data through the following features:

Feature	Scenario
Tracing	Query the complete call chain, analyze time consumption (flame graph), and locate error nodes.
Error Tracing	Aggregate and classify errors, trace error chains, and analyze error stacks.
Service Analysis	Monitor key service metrics (throughput/latency/error rate) and drill down to analyze issues.
Dashboard	Customize visual performance metrics (e.g., P90/P99 latency, error rate, call volume).
Log Correlation	One-click correlation of chain/error context logs for in-depth root cause analysis.

Through the above features, you can comprehensively grasp application performance and quickly locate issues.