cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Want to know WAN causing any delay

karthikayini_mu
Inactive

Hi All,

Hi Team,
I am working on some critical thing to find out delay acceding the use cases.
The users are in (U.S) The application Servers are in (Europe)
The users facing delay in response.
When I fetch data from user actions dashlet, then added up all the Server Contribution [s]

Got the client side by = Total -(Server+N/W)

Looks like network contribution is very less 8%
please refer the Excel Sheet
So Does it mean that WAN connection is not the causing much delay.
Need to work on the server side.

Added the monitoring window screen shot.
In there also
Desk top browser 7.35%
Weblogic xxx1.66%
Weblogic yyy0.13%
Weblogic zzz17.19%
ME15_aaa.06%
ME15_bb 4.27%
ME15_ccc 7.71%
Database 25.98% 0.13%
Oracle 6.26%
Total -99.74
What about network contribution here.How to find it from seeing it.

My main cooncern is to find whether the WAN causing any delay.

Thanks & regards

Karthikayini M

4 REPLIES 4

Joe_Hoffman
Dynatrace Champion
Dynatrace Champion

Karthikayini,,

In your Transaction flow, notice the "Inter Tier Time Per Transaction" at the top right corner. This indicates your network time is very small (.28%), so I would focus on the other components, not network. You can select "Show" to see which network segments contribute, but the total contribution isn't relevant from a performance perspective.

hth

joe

Hi @Joseph M. H.

Thanks for your reply.

Can I consider it as WAN is not causing any delay. There is no use in shifting physical servers near to user location.

Having a doubt by reading the below terms.

N/W bandwidth influences the pure path .

How to calculate only WAN contribution in terms of time for the over all transactions.

The User Action PurePaths dashlet calculates how much of the response time is split between the client, server and network time. The algorithm to calculate the three fractions works in the following way:


  1. Server time or server contribution: The sum of the calls to blocking APIs in server-side PurePaths, considering overlaps. Simply summing up the PurePaths durations (even considering overlaps) creates a wrong impression, since the network bandwidth influences the PurePath: Blocking writes on the server-side prolong the PurePath. A client's (browser's) slow network connection can prolong the PurePath.
  2. Network time or network contribution: The sum of all response sizes sent to the client / bandwidth of the client.
  3. Client time or client contribution: The rest of the user action response time is client time.

Thanks & regrads

Karthikayini m

Joe_Hoffman
Dynatrace Champion
Dynatrace Champion

Network time is calculated by taking the total user time (Browser) and subtracting the Serverside time. What is left is assumed to be network time. In your situation, would conclude that the best investment of your time would be to focus on the ServerSide. Drill down into the Method Hotspots dashlet and it will help narrow down the specific code which is most contributing to user response time problems.

Hi Joseph,

Thanks a lot for replying.

I am getting data as above in user action dashlet.Could you please help me using the above terms

Server Contribution

Network Contribution

From these data ,whether it is possible to extract only WAN contribution in the user action response time. I am bit confused in understanding the exact WAN time which is between countries.

----------------------------------------------------------------------------------------------

I am getting method hot spot as below.

Could you please help me how to find the time taking method.

Should I catch the highest execution time method or CPU time high method or High Synchronization method or high wait time method.

I generally consider highest execution time method .

Thanks you so much in advance.