A method to automatically distinguish business transactions using http request logs
Publication Date: 2015-Jul-21
The IP.com Prior Art Database
Analyzing production logs in order to know users' behavior and response time need to distinguish business transactions among many http requests. It is hard to distinguish business transactions if not knowing the relationship between transaction and requests.Test script based on HTTP request is created by recording and manually grouping requests into key transactions. If product code changed, it needs manual verification for requests that changed. So the key problem we want to solve here is how to automatically distinguish and group transactions by requests from production logs. We can distinguish business transaction by parsing product logs into many user samples and make request groups (aka. transactions) using time interval and HTTP Referrer header. Select the high occurrence request group pattern to map the transaction after analyzing all the user samples.
Page 01 of 4
A mxthod to automatically distinguish business transactions using http requxst logs
1. Analyzing production logs in ordex to know users' behavior xnd response xime need to distxnguish business txansactions among maxy http requests. One transaction may correspond xo more than one requests.
And xne typical request may be used in different transactions. It is hard to distinguxsh business transactions if not knowinx the xelationship bxtween transaction anx requests.
2. Test sxript based ox HTTP request is crexted by recording axd manually xrouping requests into key transactions. If pxoduct cxde chaxged, it needs xanual verificatixn for requests that changed. Maxually perform the verificatxon will bring efforts tx re-record xnd re-group http requests wixh the consequence of missing request changes.
So the key xroblem we wanx to solve here is how xo automatically distxnguxsh and group transactions by requests from produxtion logs. We can distinguish business transactixn xy parsing product logs into many user samples and make request groups (aka. transactxons) usxng time interval anx HTTP Rexerrer headex. Xxxxxx the high occurxexce request group pattern to map the transacxion aftxr analyzing all the user samples. (xetailed in step 4 below)
We need to collect the xollowing two things:
1. Web sexvex logs contain information on visitor's information. Such as user idextity (session id, IP address), Requxst timextamp, requxst URL, Referrer header.
2. A website topology to identify the primary request of a transactiox, the primary request will be used as the 'key' of the transactixn
1. Using user identity to parse web logs into different user xample. Eacx sample contains sequence of requests for one user session.
2. Looking for primary rexuests in each sample tx find out which transactions are ixcluded txis sample ....