Distributed - Tracing

Data System Architecture

About

tracing in a distributed world.

Implementation

Thread local storage

Instrumenting asynchronous application for distributed tracing is quite challenging because most tracing libraries rely on thread local storage (Storage that is local the thread). While it works reasonably well in a blocking application (one thread), this does not work for an asynchronous application.

This implementation supposes that:

  • the application control flow matters (i.e threads)
  • although what really matters is the application request flow (e.g the incoming HTTP request).

Library





Discover More
Data System Architecture
Conccurency - Asynchronous Model

Asynchronous allows an application to issue multiple requests and continue executing while the server performs the request. This type of request can improve an application’s throughput because it allows...
Debugging
Debugging - Log Statement

in debugging. Inserting log statements into code is a low-tech method for debugging it. It may also be the only way because debuggers are not always available or applicable. This is usually the case...
Data System Architecture
Log - Logging

A log is a text file with: a timed list of message (activities) that an application has performed execution/request log: web server: web log ( - that stores the Http request error cron...
What is a UUID? - Universally Unique IDentifier - (also known as GUID, Globally Unique IDentifier)

UUID (Universally or Global Unique IDentifier) are generated identifiers that are guaranteed to be unique and avoid collision



Share this page:
Follow us:
Task Runner