All Products
Search
Document Center

What is Tracing Analysis

Last Updated: Sep 17, 2018

Product overview

With invocation trace mapping, invocation request counting, trace topology discovery, application dependency analysis and more tools dedicated for distributed applications, Tracing Analysis allows developers to quickly identify the performance bottleneck of distributed application architecture, and significantly improve the development and diagnosis efficiency in the framework of micro-services.

Product architecture

The product architecture of Tracing Analysis is as shown in the figure.

This is how it works:

  1. The customer application reports service invocation data by integrating the multi-language client SDK of Tracing Analysis. Tracing Analysis supports multiple SDK of open-source community, and supports OpenTracing standard.
  2. When the data is reported to the Tracing Analysis console, Tracing Analysis components fulfill real-time aggregation calculation and data persistence, to get the trace breakdown, performance summary, real-time topology, and more monitoring data. You can then perform diagnosis and troubleshooting based on these data.
  3. The trace data can be imported to a variety of Alibaba Cloud products, including LogSearch, CloudMonitor, MaxCompute, which can be helpful for making offline analysis, setting alarms, or other purposes.

Product features

Tracing Analysis is capable of:

  • Querying and diagnosing distributed invocation trace: Tracking all user requests for micro-services in the distributed architecture, and combining them into distributed invocation traces.
  • Summarizing application performance in real time: Summarizing the single services and resources that make up of the application by tracking the user requests in the whole application.
  • Dynamically discovering distributed topology: Collecting the distributed invocation information from your distributed micro-service applications and PaaS products.
  • Supporting multiple programming languages: Based on OpenTracing standard, fully compatible with open-source community, such as Jaeger and Zipkin.
  • Diverse possibilities of integration with other products: Using the collected trace information directly for log analysis or integrating with downstream analysis products such as MaxCompute.