Updated on 2022-04-28 GMT+08:00

Collection Configuration

To reduce memory, database, and disk space usage, you can implement collection configuration as required. The collection configuration takes effect for selected applications.

Procedure

  1. Log in to the APM. In the navigation pane, choose Agent > Configuration.
  2. Select an application from the Application drop-down list.
  3. Click to enable data collection.

    This function is enabled by default. When you do not need to collect tracing and topology data of a specific application, disable this function to reduce resource usage.

  4. Click to enable the function of collecting normal call chain data.

    To reduce the resources consumed by probes, APM collects one more data record every minute when a transaction is abnormal or the latency is greater than Application Performance Index (Apdex) Threshold. If this function is enabled, normal call chain data is sampled and collected. If this function is disabled, normal call chain data is not collected.

  5. Click to enable memory monitoring.

    To prevent probes from affecting service performance in peak hours, enable memory monitoring. When the instance memory usage is excessively high, probes enter the hibernation state. You can also click to set the duration and memory usage.
    • Memory usage = Used memory of the Java process/Maximum available memory
    • Maximum available memory: Use the smaller value between the available memory quota of the container and the maximum heap memory of the JVM. The maximum heap memory of the JVM is the value of -Xmx. The default value is 25% of the maximum available memory of the JVM.
    • The memory usage during collection suspension must be greater than or equal to that during collection restoration.

  6. Click to enable the function of adding trace IDs to logs.

    A trace ID uniquely identifies a tracing. When this function is enabled, the system adds trace IDs to logs. You can accurately search for logs based on trace IDs, such as ffffffffe1c08cab, ffffffffe1c08cad, and ffffffffe1c08cae, as shown in the following figure.

  7. Click to enable SQL analysis.

    When this function is disabled, no SQL data is affected, but you cannot implement SQL analysis.

  8. Set the HTTP response codes to be ignored.

    To prevent unnecessary error reporting, and quickly and accurately locate faulty tracing, set the HTTP response codes to be ignored. They will not be recorded in error record tables. Click , enter the HTTP response codes to be ignored, and click . If there are multiple HTTP response codes, separate them by commas (,).

  9. Set the errors and exceptions to be ignored.

    To prevent unnecessary error reporting, and quickly and accurately locate faulty tracing, set the errors and exceptions to be ignored. They will not be recorded in error record tables. Click , enter the errors and exceptions to be ignored, and click . If there are multiple types of Java exceptions, separate them by commas (,). The default value is null.