MSA Analysis
Home > Select Project > Analysis > MSA Analysis
The MSA analysis function analyzes the proportion of calls between transactions collected through the WhaTap's Multiple transaction trace function. Based on this, dependencies between microservices can be checked at the URL level rather than instance-centric basis.
For example, there are five microservices: A, B, C, D, and E. If a user calls A, it is assumed that the following three call patterns occur according to the logic.
- A→B, C calls
- A→B, C calls + then B→D call
- A→B, C calls + then B→D, E calls
If these three patterns have been performed once, MSA analysis can check the association analysis data depending on the reference URL.
For more information about multiple transaction trace, see the following.
MSA Transaction Statistics
The following lists all the transaction statistics that have callers or callees.
If you select the Detail icon of the URL to see among the statistical data above, the detailed view window appears. In the detailed view window, you can see the detailed data such as Caller callee summary, Caller callee data, and Caller callee trend for the transaction.
If you select , the detailed view window appears in full screen.
Caller Callee Summary
The Caller callee summary tab displays the dependency data between the caller and callee of the URL. At this time, numerical data whose maximum is 1 such as 0.004 and 0.003, appears together, which indicates the proportion of all callers. Likewise, the gravities for each callee appear on the right.
Through the above screen, you can analyze the dependencies between callers and callees based on the /account/save/employee/seoul transaction.
The caller or callee nodes can be expanded into sub-nodes as follows:
Caller & Callee Data
The correlations between callers and callees for the reference URL (e.g. /account/save/employee/seoul) can be displayed in a table as follows.
Caller Callee Trend
The reference URL (/account/save/employee/seoul) and the trend of calls between a caller or a callee can be analyzed in time series as follows.