Locating the Causes of Request Errors

Background

When the number of external requests increases sharply or the load changes abruptly, application performance problems occur frequently, for example, requests cannot be quickly responded or properly handled. Quickly identifying and locating application performance problems are common in routine O&M.

APM, as a cloud application performance diagnosis service, provides powerful analysis tools. Topologies and traces show the application status and call process, helping you quickly locate performance bottlenecks.

For example, you can view the call relationships between services and quickly locate abnormal instances through topologies. You can also drill down to services and determine root causes based on method tracing.

Applicable Scenarios

• Routine inspection, covering application metrics such as latency, throughput, and number of errors
• Quick locating of error calls

Locating the Causes of Request Errors Based on Traces

1. Log in to the APM console.
2. In the navigation pane, choose Application Monitoring > Metrics.
3. Click the URL tab. On the page that is displayed, check metrics such as the number of calls, number of errors, and latency.
   Figure 1 Going to the URL page
   

4. Click the abnormal URL to go to the tracing page.
   Figure 2 URL details
   

5. Locate error or slow traces.
   Figure 3 Viewing traces
   

6. Click the corresponding URL to obtain the trace details and determine the root cause.
   Figure 4 Trace details