WebDynpro SSR / Browser Response Time

Good morning,
When we are visualizing a WebDynpro view we take an unacceptable response time (of almost 1 minute) and the CPU of the computer client rises almost until the 100%.
The View is composed by a menu to the left (which is a embedded view )and a main view, which is formed by a group that contains a Table within a ScrollContainer. So, the view is not much complex.
The table is mapped to a simple structure whose attributes are simple objects (string) and the maximum table record size is 100.
Additionally whenever any event takes place, either or in the menu of the left or the own table, the response time remains in 1 minute although business logic is not executed.
We have proven to delete the ScrollContainer and show the table but the performance doesnt improve. We have also tested that communication network problems doesnt exist.
The performance of the client browser has been verified including the SSR parameter (sap.session.ssr.showInfo=true). A document with an
image is attached, it is possible to see that the browser response time is 45 seconds to display a content of 1 MB (isn´t it too much?? Why WD generates too much HTML??).
SAP WAS 6.40 y SP15
Browser:Internet Explorer 6.0.2800.1106 SP1
Thanks in advance,
Eloy

Hi Eloy,
We also faced a similar problem in our project. When the page size reaches 0.5MB+ the reposonse becomes too slow.
This is becuase WebDynpro gets marshalled data from backend and unmarshlles it based on your screen design. So in your case if you have 100 rows * 50 columns it will unmarshall all these records at front-end i.e, the client. Hence you see the response time of your CPU reaching 100 %
You have very few options
1) Decrease the no of visible rows on the screen at a time. Say max 10. If you have 40-50 columns explore using Tab Strips with 12-15 columns in each tab.
2) Increase the RAM and Processing capabilities of your Client PC's. We were kind of lucky that our customer agreed to this and got P4 1GB machines.
Lets hope the performance is improved in the future releases.
Regards,
Shubham

Similar Messages

Browser response time slow but Analyzer fast with same query

Hi all,
We have a very complex query (structures and cell formulas). When I drill and filter on a specific characteristic the action takes anything between 7 and 15 minutes when using the browser. The same query in Analyzer performing the same drilling action takes less than 35 seconds.
Our architecture is as follows:
- Relevant cubes are indexed on BWA
- WebDynPro for Java
We can see that the BWA is hit for a short time then the CPU utilization is back to 0%. When doing '&profiling=x' on the browser I can see that the most of the time is spent on 'Get result set' / 'Get provider result set' for the Java events and on the OLAP side most time is spent on the 'Get result set' event.
Interestingly enough - when disabling myself as a user for BIA use (NO_BIA_USE) in my user profile I get the same response time on the browser.
So, does anyone have an idea of what may be the cause of the slow response time when using the browser?
Thanks
Edited by: Cobus van Rooyen on May 10, 2010 8:49 PM

Hi,
The reason is most likely because of the event ID 3200 where data transfer takes place from OLAP to frontend browser.
Record a RSTT trace for this and if you are able to see frequent calls to "Apply State XML" and "Get Result Set" then apply the SAP Note 1428850. Otherwise see which function module is causing an error.
If trace runtime is less but in browser still it takes long time then this might be because of the compression of HTPP responses, which can be confirmed by observing the sent and received data volumns from the HTTP server requests/responses. For more information refer SAP Note 746666.
Imran...

HTTP response time - From any SAP table.

Hi,
Can we get http response time other than from SMICM u2192 Http log?
Reason for me to ask is we cannot see complete URL in smicm log with response time.
date - "POST /sap/bc/webdynpro/sap/cprojects/?sap-contextid=......................................... HTTP/1.1" 200 325894 9555 h[-]
As per sap help link complete URL after contextID is hidden for security reason.
http://help.sap.com/saphelp_crm70/helpdata/EN/48/442541e0804bb8e10000000a42189b/frameset.htm
u201CFor security reasons the following information is hidden (replaced with points) from the logging procedureu201D
Question: can we access complete URL from any table level with http response time, any input is appreciated.
Thanks,
Venkat.

Michael,
First of all, thanks for the reply.
In our environment we have cprojects implemented and have lot of webdynpro request to ECC system.
From http log (SMICM) I can see many webdynpro requests with response time above 30 minutes; I want to automate this monitoring by writing a CCMS data supplier. We already have SAP document to write data supplier.
Scenario: For example if they are doing a task in cprojects.
We will be having a GET webdynpro request and then POST. By looking at GET we can know the task user is trying to do. But it is POST request which will be taking most of the response time and POST URL is show u2026(point) after context ID and as per SAP help this is done for security reason.
HTTP log from SMICMu2026
[DATE] - "GET /sap/bc/webdynpro/sap/cprojects?sap-client=clnt&STARTVIEW=Tasks&OBJECT_TYPE=TTO&GUID=GUIDNUMBER
[DATE] - "GET /sap/bc/webdynpro/sap/cprojects/~ucfLOADING?sap-contextid=....................
[DATE] - "POST /sap/bc/webdynpro/sap/cprojects/?sap-contextid=........................[response time]
I think SAP is displaying as point in http log file, but it should definitely have complete URL is any of the table.
So I am looking If we can find the complete POST URL with response time from table then we can write an ABAP program to send the information to CCMS.
Thanks,
Venkat.

Abap WebDynpro Response time monitoring.

Hi,
We are trying to monitor response time for Abap WebDynpro.
From SMICM u2192 goto u2192 Http log u2192 server display entries. We can see the response time as 43 milliseconds
07/May/2010:23:03:22 -0500] - "POST XXXXXXXX HTTP/1.1" 404 2118 [43] h[-]
By defining the parameter icm/HTTP_logging_0 we can adjust the %T value, but I want to use ALM (alert management) to send an alert if the response time is more than X seconds seconds.
I thought of using log file monitoring but help.sap.com says we can define condition like X > 3 sec send an alert instead it will search for exact error message.
I want to avoid writing script using awk(unix) as I do not have OS access.
Please let me know if anyone has automated Abap webdynpro monitoring or any other thoughts is appreciated.
Thanks,
Venkat.

HI Venkat,
Try to explore this MTE (Use CCMS) , not sure you can achieve what you need ...
WDAClass
WebDynproABAPClass
You can reach here
SID\sapserver_SID_00\Web Dynpro ABAP\CpuTime
Hope it will help

Spry menu response time problems with IE

We implemented the spry vertical menu for showing the
categories of a products catalog. It has almost 1400 categories
organizad at about 5 levels, some categories have about 20
subcategories. These categories are in a coldfusion session
variable.
It works perfect in mozilla, but in IE7 and IE6 in some
computers, present this problem:
- The response time is slow when you change from one category
that has subcategories to ahother. If you see the Windows Task
Manager of the computer while you use the menu, the processor use
go up to the top level.
And the effiecience of the menu decrease.
See in
http://edit.panamericana.com.co/
Thanks,
Alejandro

mdr4win wrote:
i dont think you understood my question, wasnt about body background, but abut the spry image slideshow to work properly in IE
I was not talking about body background, but about having markup that screws up your document when using a browser. Body background just happened to be there. Perhaps you would do well to have a look here http://validator.w3.org/check?verbose=1&uri=http%3A%2F%2Flittletreats.org%2F.
I have noticed that you ignored my solution; your document still shows two bodies.
Perhaps I should have mentioned that I tested in IE6 through to IE9 using IETester and the above was the only thing stopping IE from performing properly.
How did you determine that the slideshow was not working correctly in IE and which versions of IE?
Grumps

Slow response time for HTTP requests

I have an 8310 Pearl. For the past week, I've noticed that the response time using the Internet browser was very slow. One work-around was to cancel the request and then submit it again. The second request returned the page MUCH faster.
I cleared my history, and this seemed to have helped. Has anyone else noticed this? Thanks

Welcome to the forums
Try a Battery Pull:
1. Remove the battery while phone is turned on
2. Wait a few seconds before inserting it again
3. Allow the long boot up process to finish (takes a few minutes usually)
This clears most software issues that occur on blackberries from time to time. You could also use Quickpull, a free app that simulates a battery pull without physically removing the battery: http://software.crackberry.com/product.asp?id=27660
A battery pull reclaims memory lost during application usage and clears minor software errors that creep in over time. The blackberry runs an OS just like your PC runs windows and the battery pull is the equivalent of a reboot/restart on your computer.
Kijana
Please remember to:
1. Mark Accept as Solution on the appropriate post once your issue has been resolved
2. Give Kudos to helpful posts (click the star next to the post)
Thanks

ISA Server 2006 + Average response time for Non Cached requests = performance issues?!?!?!

All,
I am in a predicament with internet browsing speeds...We have a 3rd party look after our line and internet facing f/w so I cant troubleshoot them, so at the moment Im looking at ISA as the potential bottleneck - we have a fairly standard environment:
Internal > Local Host > Perimiter n/work > Firewall > Internet
I have been running custom reports on the ISA server to see what data can be collected - I have noticed that "Average response time for non cached requests" (traffic by time of day) can be as high as 76 seconds!!!!!! Cached hits are between .5
and 2 seconds.
I have also coonfigured a connectivity verifier which is also flagging slow connectivity, massively over the >5000ms and also reporting "cant resolve server name on occassions- and this is configured for
www.Microsoft.com --- DNS ???!?!, however I have looked through DNS (no obvious errors / config issues) which I can see
I have run the BPA on ISA server to ensure its Health - - connectivity verifier errors flagged timeouts to microsoft.com as expected...
Can anyone advise any obvious areas to investigate as Im struggling! - as always the 3rd party have told us the internet pipe is fine :O

Problem resolved.
DNS forwarders have been changed on the ISA server / DNS and this has improved lookup speed considerably.
thanks all :)

Slow response time for radio buttons on Internet Explorer 7

Hi,
We're using ADF 11.1.1.4 on weblogic 10.3.4 on a Solaris 10 (sparc) platform. The architecture also uses OHS and Webcache 11.1.1.4.
One of our applications has a webpage that is composed of several tabs, inside each tab we have some questions with radio buttons. When a client responds all questions on the first tab it moves to the second, where the application loads the necessary questions, based on the answers provided on the first tab. Once the client finishes answering this tab it moves to the third one and so on.
The problem appears once the tabs are loaded. Each time the client clicks one of the radio buttons, Internet Explorer takes between 3 to 5 seconds to return control, and considering that a client might need to click dozens of radio buttons to complete a process, it becomes a very exhausting task. It's worth noting that the session information is persisted for each radio button, when we move between tabs all answers are kept even if we didn't saved the progress.
On the other side, if we use Google Chrome 16, we find that the webpage returns control to the user almost inmediately, when clicking the radio buttons the response time is usually at 1 second or less.
We've checked that the app server is not stressed, heap memory is nowhere near the maximum assigned and the used processor % is very low, so it does not seems to be a resource problem.
Is there any known issue with ADF 11.1.1.4 and IE7/8 that affects the browser's response time? Are there any guidelines or recommendations regarding how to make the radio button component work properly? Has anybody noticed this behavior in adf?
Thanks.

I think that the most accurate description of the problem is that the autosubmit action of the radio button takes considerably longer in IE7 than in Chrome.

CCMS Monitoring for Average HTTP response time

Hi guys ,
Is it possible to monitor Av. response time for HTTP(this is based on ST03) using CCMS ?
Thanks in advance
George

hi
As per my understanding you tried to view the responsetime comp as same like in sto3 by ccms, the way as below, RZ20-> SAP CCMS montior templates- >Entiresystem-><sid>->application server -><server>R3 services -> DIalog you can get all the response time as same in the ST03.
else in otherway if you looking for monitoring Java and webdynpro, you can do this by GRMG.chk this for to knw http://help.sap.com/saphelp_nwpi71/helpdata/en/45/4ee37fea5917eae10000000a155369/content.htm|http://help.sap.com/saphelp_nwpi71/helpdata/en/45/4ee37fea5917eae10000000a155369/content.htm]
for configuration,[GRMG-config guide|http://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/106a4d88-2a36-2c10-549b-dc549d00b8b9?quicklink=index&overridelayout=true]
chk and update
jansi

What exactly does the contrast ratio and response time mean?

Recently I have been browsing for external monitors and I noticed that two of the specifications for monitors are a contrast ratio and response time.
What does a contrast ratio mean? What does the response time mean? Is the response time the same as the frequency refresh rate (ie. 75 Hz)?
I was looking at the Thunderbolt Display specs and it has a contrast ratio of 1000:1 and a response time of 13ms, but another LED monitor that costs 1/3 the price has a contrast ratio of 5,000,000:1 and response time of 5ms. The resolution of the display does not matter too much for me. Would the cheaper monitor be a better choice for me if I play to mostly use the display for gaming/photo editing?

You're welcome.
Would the cheaper monitor be a better choice for me if I play to mostly use the display for gaming/photo editing?
You can problem find lots of decent monitors with short response times for gaming. But I think (and this is just a guess, no basis in fact) that a decent monitor for photo editing may not be cheap and of course response time doesn't much matter so it may not make a good game monitor (depending on the kind of games you like to play of course -- angry birds would play on anything).
Also another thing to keep in mind are the two schools of thoughts on glossy monitors like the current 27" Apple monitors vs. non-glossy (anti-reflective) monitors. It's like religiion, And one group is not going to convience the other which is "better". It's a matter of taste I think. There's even a thread on this topic in these forums:
Why is Apple insisting on Glossy Displays?
I'm only mentioning this because if you want to do some photo editing there are some folks who do that stuff and claim they don't like glossy displays.

Webgui and high response time

Hi guys,
we have upgraded our worktation to windows 7 with IE9. Our users complains high response time after migration.
We have tested the webgui with Win 7 and IE8 and the response time is accetable, but we can't downgrade the IE version.
We have tried to setup the compatibility with IE8 , but the issue persist.
Do you have any suggestion about resolve the issue ?
thank's in advance fro your help.
Antonino

What is your SAP MII version?
While loading webpage, is it showing contentiously loading and not opening content, then its 100% java issue only.
Goto IE9-->Tools-->internet options --> Advanced --> Browsing check(Disable script debugging). Once you down this, on status bar its shows the error like 'undefined object'.
To correct it, you need to upgrade or degrade java version accordingly.
Mine is Window 7 and using Java version 1.7.0_13-b20. its working fine.
Regards,
Praveen Reddy

Problem with response time

Hi, I am experiencing long response time for my dashboard(40 seconds on some PC). I wonder what is the normal response time for your dashboards, and what is the size of your worksheet?
For me, I have a tab set with 5 tabs, under each tab, there are about 100 row* AH cloumn. And as it is still the proof of concept stage in my company, I am using offline data, but if we use live data from BW, what would it mean to the resopnse time, even longer?
Thank you very much.

Hello,
We try to keep our SWF size under 1024 kb, but that doesn't mean it will load fast since the connection loads little after Initializing the component. We have around 100 components on an average, and some custom components. One thing you should stay away from is doing lot of grouping in the Object Browser.
Questions for you:
Which version of xcelsius and service pack are you using?
How many connections run when the dashboard loads initially?
How many total connections?
How many rows and columns does the initial load connection bring?
How much time does the same query take in the database?
Edited by: msaraogi on Dec 15, 2010 10:32 PM

Avg iview response time

At Workload Analysis (Sol man Diagnostics) why I have u201CAvg servlet/webdynpro response timeu201D and u201CAvg iview response timeu201D?
I mean, the request time of the iview is calculated separated from the request time of the servlet (In case I have an iview with a servlet inside), if so I have to sum both to have the complete time request? Or the KPI u201Cavg iviewu201D is for the iviews that does not have a servlets or a webdynpros inside?
I am a little confused with these concepts, someone can explain to me or show me some document that explains this?
Thanks in advance
Regards,
Walde Requena

I don't understand the question.
Do you want to know how to display dialog response time in ST03? If so, your instructions tell how to bring up that screen. It will show Dialog response time providing there have been Dialog steps executed.
Or do you want to know which function module you call to get dialog response time data? Use SAPWL_WORKLOAD_GET_STATISTIC to get hourly, daily, etc. data (for a given day and instance) for each task type.
Tim

Response time & dump-allowed & disable connection pooling

3 Questions:
Question 1:
Where do I find status report when I set "dump-allowed" to yes? Where does it get "dumped"? I've looked in servlet log files, such as:
1. C:\Program Files\New Atlanta\ServletExec ISAPI\Servlet Logs\Servlet.Log
C:\WINNT\SYSTEM32\LogFiles\W3SVC1\exyymmdd.log
in a NewAtlanta ServletExec 4.1 for IIS/PWS 5.1 scenario AND
2. C:\oracle\ora81\Apache\Jserv\logs\Jserv.log
C:\oracle\ora81\Apache\Apache\logs\access_log
in a Apache 1.3.12 with JServ 1.1 scenario (the versions that come with Oracle 8.1.7 on W2K).
but cannot find any connection pool dump information in any of the above log files.
Is there documentation on "dump-allowed" other than the short blurb in the XSQLConfig.xml file?
Question 2:
Does setting the following connection pool elements, initial & increment, in the XSQLConfig.xml file
<initial>1</initial>
<increment>1</increment>
to 1 & 1 respectively, actually DISABLE connection pooling?
When I set the timing-info element to "yes", I see about 2 to 4 second response time to any of the Insurance demos at the outset. This is probably normal to establish the first connection and to process the SQL statement. However after 60 seconds, which is the setting for the timeout-seconds element, I again see about 2 to 4 second response time. I expect a response time under 1 second because not only should the SQL statement be cached since it had previously been parsed, the initial connection should still be live.
Thus it seems that an initial setting of 1 does not do what it's suppose to do, and that is maintain a connection pool of 1.
I thought I had read in the XML forum that:
<initial>1</initial>
<increment>0</increment>
does disable connection pooling. Not sure how and why. What does disabling connection pooling really mean anyway?
BTW I'm using XDK 9.2.0.1.
Question 3:
The real reason behind questions 1 & 2. How do I track response times from end to end?
I know about setting timing-info element to "yes" in the XSQLConfig.xml file.
<timing-info>
<page>yes</page>
<action>yes</action>
</timing-info>
However if I want to know where a hold up is, timing-info is not enough.
How do I turn on sql trace for the session affiliated each JDBC connection. I know about setting time_statistics= true in the init.ora file, but how do I alter session set sql_trace=true for an XSQL query action?
I think I answered this question. See following amended claim1.xsql file for the Insurance example:
Claim1.xsql:
<?xml version="1.0"?>
<page connection="demo" xmlns:xsql="urn:oracle-xsql">
<xsql:dml><![CDATA[
ALTER SESSION SET SQL_TRACE = true
]]></xsql:dml>
<xsql:query><![CDATA[
select value(c) as Claim from insurance_claim_view c
where c.claimpolicy.primaryinsured.lastname = 'Astoria'
]]></xsql:query>
<xsql:dml><![CDATA[
ALTER SESSION SET SQL_TRACE = false
]]></xsql:dml>
</page>
The result of which is:
<?xml version="1.0" ?>
- <page xsql-timing="3485">

<xsql-status action="xsql:dml" rows="0" />

- <ROWSET>
- <ROW num="1">
- <CLAIM>
<CLAIMID>77804</CLAIMID>
<FILED>1/1/1999 0:0:0</FILED>
</CLAIM>
</ROW>
- <ROW num="2">
- <CLAIM>
<CLAIMID>12345</CLAIMID>
<FILED>3/11/1998 0:0:0</FILED>
</CLAIM>
</ROW>
</ROWSET>

<xsql-status action="xsql:dml" rows="0" />
</page>
and in the database trace file:
select value(c) as Claim
from insurance_claim_view c
where c.claimpolicy.primaryinsured.lastname = 'Astoria'
call count cpu elapsed disk query current rows
Parse 1 0.00 0.08 0 0 0 0
Execute 1 0.00 0.07 0 0 0 0
Fetch 1 0.00 0.21 0 11 12 2
total 3 0.00 0.36 0 11 12 2
Do I use the elapsed time to understand the time spent in the db? Do I consider the other SQL statements in the trace file that seem to deal with the page request? If so, then there is an overall timed total provided.
OVERALL TOTALS FOR ALL NON-RECURSIVE STATEMENTS
call count cpu elapsed disk query current rows
Parse 14 0.00 0.20 0 0 0 0
Execute 15 0.04 1.07 0 14 0 6
Fetch 7 0.00 0.38 0 210 12 21
total 36 0.04 1.65 0 224 12 27
Misses in library cache during parse: 0
OVERALL TOTALS FOR ALL RECURSIVE STATEMENTS
call count cpu elapsed disk query current rows
Parse 2 0.00 0.16 0 0 0 0
Execute 7 0.00 0.04 0 0 0 0
Fetch 7 0.04 0.25 0 56 0 7
total 16 0.04 0.45 0 56 0 7
Misses in library cache during parse: 0
So do I use overall elapsed time in db affiliated with the Claim SQL statement, which is 2.1 seconds? Is the 3.5 seconds in the XSQL timing an elapsed time, and does it include time spent (2.1 seconds) in the database?
I guess you can see where this is going. What I really need is a recommended plan to analyze response time problems.
Has anyone come across this respose time problem, and how did you attack the problem? Did you use timing-info and cross check with database timed statistics using tkprof? What database statistics did you use?
How does one analyze response time even further. For example how does one track time within the web server? How does one track time within the servlet engine?
Ideas anyone?
Steve.
The section of the XSQLConfig.xml file to which I refer:


[list=1]
[*]Where do I find status report when I set "dump-allowed" to yes?
The report is returned in the browser.
[*]Does setting the following connection pool elements, initial=1 & increment=1 actually DISABLE connection pooling?
No. This tells the pool to start a one connection, and to grow by one connection whenever the pool finds all current connections in-use in the pool.
When I set the timing-info element to "yes", I see about 2 to 4 second response time to any of the Insurance demos at the outset.
Initial = 1 means that the pool will start with 1 connection. It also means the pool will never shrink to below 1 connection.
I thought I had read in the XML forum that initial=1, increment=0 disables connection pooling.
Setting increment=0 forces the pool never to grow, but if you have more than one user this is not a desireable situation to be in. It will cause potential delays for users contending for the one connection available.
[*]How do I track response times from end to end?
<timing-info> is the only XSQL pages specific timing info available. You saw how to use <xsql:dml> to turn on tracing. Once you have the trace output, the analysis is a database profiling task.
[list]
You can experiment with the fetch-size parameter to see if increasing the number of records fetched at a time will improve your throughput.

Urgent:How to Speed Up the Response time for PDF Report of 500 pages

hi all
i am running the 9ias on Solaris and Generating some reports
Which fetches around 50,000 records and Display them in pdf format.
When i am ruuning the Query at Database level Response time is Very fast.
But when i Run the Report in Web Browser same report takes 7-8 minutes.
So it seems that the Conversion to pdf and Displaying it takes most of time.
Does any one has idea What parameter's needs to be changed/Caching ??? /
or any other ways or methods which can be used to Reduce the Response time.
(its once a monthly Report and User need to Download all the 500 pages as
a Single Document).
Any help / suggestion please
thanks
jai

You aren't by any chance calling a function in your repeating frame that in turn goes back and queries the database, are you? If so ... don't. We regularly do 500+ page PDF-file reports, and one thing we discovered early on was that repeatedly going back to the database while generating the report output (in our case, in calculations that were being done on each line of a report) slowed the output down by an order of magnitude. Instead, we now retrieve all the data needed for each report up front (via functions or views called in the initial SQL for the report), and just use Reports to format the output. MUUUUUUUCH faster -- 200 page reports that used to take 15 minutes to complete now complete in just seconds.
One way you can visually see if this is part of your problem is to watch the report execute in the Report Queue Manager application. If it spends all its time on the "Opening" stage then breezes through each page, this is not your problem. If instead it seems to take a long time generating each page, I'd suspect that this may be at least part of your delay.
- Bill

WebDynpro SSR / Browser Response Time

Similar Messages

Maybe you are looking for