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, locating, and handling these problems are required in routine inspection.

APM has powerful analysis tools for cloud application diagnosis. It displays application statuses, call processes, and user operations through topologies and tracing, so that you can quickly locate and resolve faults and 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

Procedure

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, view metrics such as the number of calls, number of errors, and latency.
   Figure 1 Viewing URLs
   

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