AWS Knowledge Hub

English

Tiếng Việt

English

Tiếng Việt

Appearance

AWS X-Ray ​

AWS X-Ray helps developers analyze and debug distributed applications, including those built using a microservices architecture.

Overview ​

X-Ray provides a complete view of requests as they travel through your application, showing a map of your application's underlying components and helping you identify performance bottlenecks and errors.

Key Features ​

  • Distributed Tracing: End-to-end request tracing
  • Service Map: Visualize application architecture
  • Performance Analysis: Identify bottlenecks
  • Error Tracking: Find and debug errors
  • Sampling: Control trace volume
  • Annotations: Add custom metadata
  • Filtering: Filter traces by criteria
  • Integration: Works with many AWS services

Core Concepts ​

Traces ​

End-to-end request paths through your application. A trace consists of segments and subsegments.

Segments ​

Represent work done by a single service. Contain:

  • Timing information
  • Request/response data
  • Metadata
  • Subsegments

Subsegments ​

Represent work done within a service, such as:

  • Database queries
  • HTTP calls
  • External service calls

Service Map ​

Visual representation of your application architecture showing:

  • Services and their relationships
  • Request flow
  • Error rates
  • Latency

How It Works ​

Instrumentation ​

  1. SDK: Instrument your application with X-Ray SDK
  2. Sampling: X-Ray decides which requests to trace
  3. Segments: Application creates segments for work
  4. Submission: Segments sent to X-Ray service
  5. Processing: X-Ray processes and stores traces
  6. Visualization: View traces in X-Ray console

Sampling ​

  • Control which requests are traced
  • Reduce cost and overhead
  • Configurable sampling rules
  • Default: 1 request per second, 5% of additional requests

Supported Services ​

Automatic Instrumentation ​

  • API Gateway: Automatic tracing
  • Lambda: Automatic tracing
  • Elastic Beanstalk: Automatic tracing
  • ECS: Container-based tracing
  • EC2: SDK-based tracing

Manual Instrumentation ​

  • Applications: Use X-Ray SDK
  • Custom Segments: Add custom segments
  • Annotations: Add metadata
  • Metadata: Add key-value pairs

X-Ray SDK ​

Supported Languages ​

  • Java
  • Node.js
  • Python
  • .NET
  • Go
  • Ruby

Features ​

  • Automatic instrumentation
  • Manual instrumentation
  • Annotations and metadata
  • Subsegments
  • Error recording

Service Map ​

Components ​

  • Services: Your application services
  • Edges: Connections between services
  • Metrics: Error rate, latency, throughput
  • Time Range: Filter by time period

Use Cases ​

  • Understand application architecture
  • Identify dependencies
  • Find performance bottlenecks
  • Track error propagation

Trace Analysis ​

Trace Details ​

  • Complete request path
  • Timing information
  • Request/response data
  • Error information
  • Annotations and metadata

Filtering ​

  • Filter by service
  • Filter by error
  • Filter by response time
  • Filter by user
  • Custom filters

Best Practices ​

  • Enable X-Ray: For all services
  • Use Sampling: Control trace volume
  • Add Annotations: Custom metadata
  • Monitor Service Map: Track architecture
  • Set Up Alarms: For error rates
  • Use Filtering: Find specific traces
  • Instrument All Services: Complete visibility
  • Monitor Costs: X-Ray charges per trace
  • Use Annotations: For debugging
  • Regular Reviews: Analyze performance

Use Cases ​

  • Microservices Debugging: Debug distributed systems
  • Performance Optimization: Find slow operations
  • Error Tracking: Identify and fix errors
  • Architecture Understanding: Visualize application
  • Dependency Analysis: Understand service dependencies
  • Troubleshooting: Debug production issues

Integration with AWS Services ​

API Gateway ​

  • Automatic tracing
  • Request/response data
  • Error tracking

Lambda ​

  • Automatic tracing
  • Function execution time
  • Error tracking

ECS/EKS ​

  • Container-based tracing
  • Service discovery
  • Load balancer integration

CloudWatch ​

  • Metrics from traces
  • Alarms on error rates
  • Log integration

Cost Optimization ​

  • Use sampling to control trace volume
  • Set appropriate sampling rules
  • Monitor X-Ray costs
  • Use filtering to reduce analysis time
  • Delete old traces (automatic after 30 days)

Limits ​

  • Traces per Second: Unlimited (with sampling)
  • Trace Retention: 30 days
  • Annotations per Segment: 50
  • Metadata Size: 64 KB per segment
  • Subsegments per Segment: 1,000