Skip to content

Latest commit

 

History

History
25 lines (17 loc) · 1.3 KB

README.md

File metadata and controls

25 lines (17 loc) · 1.3 KB

Deepflow-App

DeepFlow-app is based on the DeepFlow visibility platform, providing the ability to call the link trace function

Background and premise

DeepFlow stores the details related to system calls and network calls in Log Event format, and there are three kinds of data according to the Flow type classification:

  • a single request for Event data.
  • a single response to Event data.
  • A session event data, this type of data conforms to the basic definition of Span-like data.

So, here is to convert the existing Flow Log Event into Span-like data, that is: merge the request and response data in the flow type into the seesion data, and merge and sort with the session data belonging to the same call, and convert it into a call Span-like data in stack order.

How to construct Span-like data

  • Query all related Flow Log Events based on _id.
    • Use tcp_seq to query net spans.
    • Query sys span using syscall_trace_id.
    • Query app span(APM) data with trace_id, span_id, parent_span_id, x_request_id.
  • Merge request and response data into seesion data and construct Span-like data.
  • Sort the merged data and build Span-like data that conforms to the call stack.

Others

For details, please refer to here