Error Tracking Exception Replay

이 페이지는 아직 한국어로 제공되지 않으며 번역 작업 중입니다. 번역에 관한 질문이나 의견이 있으시면 언제든지 저희에게 연락해 주십시오.
Exception Replay for APM Error Tracking is in beta.

Overview

Exception Replay in APM Error Tracking automatically captures production variable values so you can reproduce exceptions from Error Tracking issues.

Error Tracking Explorer Exception Replay

Requirements

Supported languages
Python, Java, .NET
  • Your Datadog Agent must be configured for APM.
  • Your application must be instrumented with:
    • ddtrace for Python
    • dd-trace-java for Java
    • dd-trace-dotnet for .NET

Exception Replay is only available in APM Error Tracking. Error Tracking for Logs and RUM is not supported.

Setup

  1. Install or upgrade your Agent to version 7.44.0 or higher.
  2. Ensure that you are using:
    • ddtrace version 1.16.0 or higher.
    • dd-trace-java version 1.35.0 or higher.
    • dd-trace-dotnet version 2.53.0 or higher.
  3. Set the DD_EXCEPTION_DEBUGGING_ENABLED environment variable to true to run your service with Error Tracking Exception Replay enabled.

Redacting sensitive data

By default, variable data linked to specific identifiers deemed sensitive, such as password and accessToken, are automatically redacted. See the full list of redacted identifiers.

You can also scrub variable data for PII by:

To learn more about scrubbing variable data, see Dynamic Instrumentation Sensitive Data Scrubbing.

Getting started

  1. Navigate to APM > Error Tracking.
  2. Click into any Python Error Tracking issue and scroll down to the stack trace component.
  3. Expand stack frames to examine captured variable values.

Troubleshooting

A specific Python error trace does not have variable values

To keep the performance overhead of the feature at a minimum, error capturing is rate limited: one error per second includes variable data. If you don’t see variable values on a given trace:

  1. Click View Similar Errors.
  2. Expand the time range selection to find another instance of the exception where variable values were captured.

Further Reading

PREVIEWING: may/unit-testing