X-Ray

AWS _X-Ray is a AWS Managed service for diagnosing issues (performance and failures) in a microservice architecture. The main idea is that you pass headers in your calls, AWS picks up on those headers and looks at the propagation of the requests across bits of AWS.

It's a distributed tracing system.

Anatomy of X-Ray

flowchart LR
    API["X-Ray API"] --> Console["X-Ray Console"]
    Daemon["X-Ray Daemon"] --> API
    SDK["X-Ray SDK"] --> Daemon
    Other["Other Clients:\nAWS SDK or AWS CLI"] --> API
    Other --> Daemon
    API --> Other

X-Ray SDK

The SDK provides interceptors to trace HTTP requests
Client handlers to instrument AWS SDK clients that your application uses to call other AWS Services
A HTTP Client to use to instrument calls to other internal and external HTTP web services

X-Ray Instrumentation

Measure he level of product's performance to diagnose errors and write trace information.

X-Ray-Daemon

It buffers segments into a queue and uploads in batches. It's available to be installed on Linux, Windows, Mac Comes with _Elastic BeanStalk and _Lambda by default.

X-Ray Concepts

Segments:
- Details about the host (IP address/kind of host).
- Request information.
- The time/duration.
- Errors and Faults.
- Response Statuses.
Subsegments:
- More granular timing information.
- Can contain calls to other AWS services
Service Graph:
- Just the UI graph in the Console.
Traces:
- A Trace ID is the path of a request through your application.
- Gets passed along to everything else, so that it can be tracked.
- Not every trace gets recorded.
Sampling:
- _X-Ray uses a sampling algorithm to determine which requests get traced.
- Default is:
  - The first request each second...
  - And 5% of any additional requests.
- Idea is to reduce service charges when you are getting started. The default sampling rate is low.
- You can create limits for the rates
Tracing Header:
- All requests are traced up to a configurable minimum.
- After that minimum, a percentage of requests are traced to avoid unnecessary cost. (See above)
- [The sampling decision and the trace ID are added to HTTP requests in tracing headers named X-Amzn-Trace-Id.]{.underline}
- The Tracing Header determines if the trace shows up in the viewer.
Filter Expressions:
- Applications make a lot of data, you can filter it
Groups:
- Named Filter Expressions.
- Groups only update after you update them, the data before still remains (subsequent traces only)
Annotations and Metadata: *
Errors, Faults, and Exceptions: *

2.7 KiB Raw Blame History