Examkin

Question

Difficulty: HardInstrumenting Distributed Tracing with AWS X-Ray

A containerized Node.js application is deployed on Amazon ECS using the AWS Fargate launch type. The application processes incoming HTTP requests and offloads compute-heavy processing and downstream database writes to asynchronous worker threads using the Node.js worker_threads module. The application uses the AWS X-Ray SDK to capture traces. The developer has enabled the AWS X-Ray daemon in a sidecar container and initialized the SDK using AWSXRay.captureAWS(AWS). However, the application logs show 'SegmentNotFoundException: Failed to get the current sub/segment from the context' during DynamoDB writes inside the worker threads, and these downstream calls are missing from the X-Ray traces. Which action should the developer take to resolve this issue?

  1. A
    Initialize the DynamoDB client inside the worker thread by hardcoding temporary IAM credentials retrieved from a custom metadata endpoint, ensuring the worker thread has explicit write permissions for AWS X-Ray.
  2. B
    Increase the task-level CPU allocation and configure the client HTTP timeout setting in the AWS SDK configuration within the worker thread to prevent execution context expiration.
  3. Extract the current tracing header in the parent thread using the X-Ray SDK, pass it to the worker thread, and manually initialize and set the segment context in the worker thread before invoking the DynamoDB client.Answer
  4. D
    Increase the visibility timeout of the Amazon SQS queue that feeds tasks to the application to prevent the tracing context from being prematurely deleted from the queue before the worker thread executes.

Answer

Extract the current tracing header in the parent thread using the X-Ray SDK, pass it to the worker thread, and manually initialize and set the segment context in the worker thread before invoking the DynamoDB client.
The correct action is to extract the current trace header from the parent thread and pass it to the worker thread. Because the AWS X-Ray SDK for Node.js uses continuation-local storage (CLS) to track the active trace context, this context is lost when spawning a new worker thread. The developer must manually pass the trace header string, instantiate a new segment/subsegment using this header in the worker thread, and set it as the active context using the SDK.

Step-by-Step Solution

1
Retrieve the trace header from the active segment in the main parent thread using the AWS X-Ray SDK.
The parent thread obtains the trace context string containing the trace ID, parent segment ID, and sampling decision.
The main thread has access to the continuation-local storage where the incoming request's trace context is automatically tracked.
2
Pass the trace header string as worker data when spawning the worker thread.
The worker thread receives the tracing metadata upon initialization.
Spawning a new thread creates a separate execution context, meaning context tracking via continuation-local storage is not shared by default.
3
In the worker thread, initialize a new segment or subsegment using the received trace header and set it as the active context using the X-Ray SDK.
The X-Ray SDK in the worker thread now has an active segment context populated in its thread-local storage.
This allows downstream AWS SDK clients in the worker thread to successfully associate their calls with the parent trace rather than throwing a SegmentNotFoundException.

Key Concept

Manual segment context propagation across asynchronous thread boundaries in the AWS X-Ray SDK.
Rate this question

Topics

Instrumenting Distributed Tracing with AWS X-Ray

More questions from this topic