Examkin

Soru

Zorluk: ZorInstrumenting Distributed Tracing with AWS X-Ray

A developer is running a Python daemon application on an Amazon EC2 instance. The application processes tasks by retrieving messages from an Amazon SQS queue and making downstream API calls to an external gateway using the `requests` library. The AWS X-Ray daemon is running on the EC2 instance, and the EC2 instance profile has the `AWSXRayDaemonWriteAccess` policy attached. The application code imports `patch_all` from the AWS X-Ray SDK and calls it at startup. However, when the application runs, the external API calls do not appear in the X-Ray console, and the logs display `SegmentNotFoundException` errors.

What is the root cause of this issue?

  1. The hosting environment on Amazon EC2 does not automatically initialize a trace segment. The developer must manually start and end a segment in the code using the X-Ray SDK.Cevap
  2. B
    The `patch_all()` function only instruments AWS SDK clients. The developer must manually extract the trace header from the SQS message and inject it into the HTTP headers of the `requests` library call.
  3. C
    The X-Ray SDK does not support the EC2 instance profile for authentication. The developer must explicitly initialize the X-Ray recorder with hardcoded AWS access keys.
  4. D
    The SQS visibility timeout is shorter than the time taken to call the external API, which causes the message to reappear on the queue and invalidates the active trace segment.

Cevap

The hosting environment on Amazon EC2 does not automatically initialize a trace segment. The developer must manually start and end a segment in the code using the X-Ray SDK.
Unlike AWS Lambda, which automatically initializes a parent segment for every invocation, self-hosted environments such as Amazon EC2 do not automatically manage trace context. When a patched HTTP library like `requests` tries to trace an outbound call, it attempts to generate a subsegment. Because there is no active segment initialized in the thread context, the SDK throws a `SegmentNotFoundException`. To resolve this, the developer must explicitly start a segment (e.g., using `xray_recorder.begin_segment()`) and end it after processing is completed.

Adım Adım Çözüm

1
Analyze the error log containing `SegmentNotFoundException`.
Identify that the X-Ray SDK is attempting to create a subsegment for the downstream HTTP request but cannot find an active parent segment in the current execution context.
Patched libraries automatically attempt to create subsegments, which require a parent segment to exist.
2
Compare the runtime environment (Amazon EC2) behavior with AWS Lambda.
Acknowledge that AWS Lambda automatically manages the lifecycle of the trace segment (facade segment), whereas EC2 does not provide any automatic segment management for custom daemon applications.
This determines whether the framework/infrastructure or the code itself must manage the lifecycle of trace segments.
3
Determine the necessary code modification to establish the context.
Wrap the worker loop execution or downstream calls in a manually created segment using `xray_recorder.begin_segment('segment_name')` and `xray_recorder.end_segment()` or the corresponding context manager.
Explicitly declaring a segment provides the parent context required by the patched `requests` library.

Anahtar Kavram

X-Ray Segment Lifecycle Management on Non-Managed Environments
Tahmini Süre:2m 0s
Bu soruyu puanla

Konular

Instrumenting Distributed Tracing with AWS X-Ray

Bu konudan daha fazla soru