In transaction flow, I see most of the time(60-70%) is spent on the Web server but when I check the purepath tree, I don't see the drill down under web server agent.
We are trying to understand why web server is taking so much of time in serving the request.
Please find the attachment webserver1.dts and check purepaths dashlets
In the transaction flow it shows 4.57s is spent on webserver but when I see the purepath tree I see nothing for web server agent.
can you please help?
Unfortunately the (native) web server agent (Apache 2.2 in your case) does not provide the same level of deep diagnostic details as a Java or .NET agent and neither does it support placing of custom sensors AFAIK.
In your case it appears the time is spent before even making the backend call (as can be seen in the "Elapsed time" row of the PP dashlet). If the web server is communicating synchronously with other components (authentication etc.) then it could be worth checking out those components, too (maybe consider adding additional agents on related components).
Otherwise you could also check if the web server is suffering from thread pool congestion by drilling down to the host and process in the infrastructure dashboard and verifying not all threads are busy.