Average processing time too long in workload monitoring (ST03N)

Hi expert,
I wander why average processing time too long in workload monitoring (ST03N)
for example
avg.response time : 4700 (ms)
avg.processing time : 4200 (ms)
avg.CPU time : 300 (ms)
avg.DB time : 200(ms)
OS (CPU and memory ) status is nomal.
please tell me why processing time too long and what do I check more.
thanks

Hi
Processing Time is the total time taken i.e. Time when user clicks on submit button to send data to SAP System till the time SAP System processes everything and shows the updated screen back to the user. It includes Network Time also.
Average Processing Time is the Processing Time per Dialog Step
In ST03N workload Monitor, there are different Task Type like Dialog, Background, RFC, Update,..........
Could you be more specific where do you find the Average Processing Time high ? I mean for which Task Type.
If you have users complaining for Performance problem check their Network Time also.
Let us know, if this information was helpful to you.
with regards,
Parin Hariyani

Similar Messages

Profile WS run time too long

My first test run (in dev) of the custom Java PWS against a SQL server database seems to be taking too long. Its seems to be taking appx. 1 second or more per user. This will be a problem when processing around 20,000 users. We can't have a PWS running for 4-6 hours everynight.
I'd really appreciate any tips to optimize this for faster performance. I am opening DB Connection and running 2 queries for each users and clsoing the statement and connection. Can I do this another way so I open the DB connection(and maybe statement) for the Profile Source only once and release the connection etc. after the whole job is done or in case of errors.
Please help as this could be a huge bottleneck for us!
Thanks.
Vanita
Staples

Hi Akash,
Thanks for your quick reply. Unfortunately I don't have any signature field to limit the profile sync by. I am currently running the PWs in ToMCAT 4.1.30. We plan to deploy on WAS in dev. I am not sure if there is connection pooling being done by Tomcat in 4.1.30 version. Do you have any information on that? I am doing db connect and 2 property queries in the getuserproperties method.
1. Would the connection pooling (if done in WAS) make any difference to the run time?
2. Would opening the connection in initialize() method versus GetUserProperties() make any difference?
As always, thanks for your help.
Vanita
------- Akash Jain wrote on 3/1/05 1:25 PM -------Hi Vanita,On recent hardware, you should be able to perform an initial profile sync at at a rate of ~10/second. This means you should be able to perform your 20k users profile sync in under an hour. Resyncs should be much faster if you use a signature attribute.
I'm going to assume you're hitting some database backend with a table structure like the following:Users Table String UserGUID Date LastModified
Properties Table int PropID String UserGUID String PropValue
You have your users keyed off a unique name - a GUID in this example - and properties in a seperate table keyed off PropID and GUID.
Lets review the protocol and each step:a) initialize() - sends the parameters of the profile sync to the PWS, this is a good place to do a single query to your database in order to cache all user unique names and signatures (LastModified dates) in a HashTable. This will make re-syncs much faster since subsequent AttachToUser() and GetUserSignature() calls will be derived from this HashTable.b) attachToUser() - in this call you can simply lookup your user record against the HashTable created in Initialize(). If no entry is returned, then throw a NoSuchUserException. If a user does exist continue.c) getUserSignature() - again use the HashTable created in Initialize() to lookup the signature for this user. return it as a String.d) getUserProperties() - if called, this means the signature you sent back in step (c) has changed since the last profile sync. you now want to make a call to your properties table (a single DB call) to load all the property values for the user. return these as a UserPropertyInfo object.
During an initital sync, you will always get to step (d) above. During re-syncs, assuming there is low churn, I'd say a max of 1% of your calls will get to step (d) and thus the re-sync should be an order of magnitude faster in most cases.
With respect to database connections - if you are opening and closing a connection for each user, this is pretty poor with respect to performance. Your best bet in Java is to use a single connection (this is a single threaded process) which is setup in initialize(). In the shutdown() method, close this connection.
I hope this helps, the combination of using a single connection, using the signature attribute and caching all the users unique names and signatures in one call at the start of the profile sync should drastically increase performance.
Thanks,Akash

MGP Tuning: MGP Processing taking too long

Hi All,
I have question here related to MGP Processing. MGP Processing is taking too long for us, close to 900 secs. Is there a way to tune and optimize this MGP Process. Right now we have only 3 users syncing and this kind of delay in Sync for users in testing Envt is not acceptable to the client. Is there a way to tune this MGP Process. I am aware of the Consperf process. Should I be running the Consperf for all the publication items to optimize their Sync process or should it be only for those publication items that are taking too long in the Compose/Apply process.
Thanks in advance for all suggestions.

Hi Rekounas,
what would be your recommendation for some 70 snapshots and 50 users? And is there any recommendation for mobile server? We have the mobile server on dedicated machine, which is quite old piece of HW. I presume that it should have just some amount of RAM available for caching of MGP data and that it shouldn't be CPU-power consuming much - is that right? At the monent the compose phase of MGP takes some 25-30 seconds per user when there's some new data to process...

Elapsed Timer Too Long?

Is there a limit to the duration I can input into an elapsed timer VI? In my program the user can specify a duration of up to 24 hrs. My program is supposed to hold current conditions for the specified durations. When the user sets it 1, 2, 3 or four hours it holds the current conditions fine. When they set it to 24 hours the current conditions are not held for 24 hours. Is this too long of a duration for this VI (86,400 seconds)?
thanks.

rchaoua wrote:
The question regarding maximum duration a elapsed time VI has still not been answered.
You can convert the "elapsed time" express VI to a plain VI by "right-click...open front panel" and inspect the code yourself. There are no secrets! You can see that it converts the timestamp to DBL and operates on the lower precision DBL values. This should be sufficient for much more than 24 hours, so something else is wrong in your case.
Can you attach an example that shows the incorrect behavior (=premature elapsation?).
rchaoua wrote:
But I have another question; Can you have multiple elapsed time VI in a single piece of code and not have it affect the other elapsed VIs? I have three elapsed timers in my code (in different panels) and get the sneaking supsicion that they are not behaving nicely with each other. But I am not sure. I was hoping someone has expereince with this scenario and could shed some light.
The crucial subVIs are reentrant and thus there should be no interaction between multiple instances of the "elapsed time" express VI. Each will keep it's own state. Why do you suspect that they interact?
LabVIEW Champion . Do more with less code and in less time .

NI7358 Acceleration time too long

Hi all,
First, let me introduces briefly our application for your better understanding before touch the problem.
We're using NI7358 to control 8 servo motors(Mitsubishi HC-KFS13, 100W, 3000RPM, 131072Counts/Rev) for pick and place. The amplifier is also Mitsubishi one MELSERVO-J2M-P8A. From view of NI7358 all 8 axes are open-loop stepper while amplifier will close-loop control the servo motors. One limitation here is this amplifier only can accept 200KPPS pulse rate input, so we set amplifier internal gear ratio to 4096/125 to achieve its maximum speed.
GearRatio=(3000rpm*131072count/60)/(200*1000counts/sec)
Then based on this ratio we can know the resolution of "stepper" in view of NI7358 is 4000RPM. (131072*125/4096)
Now our problem is:
The system requires these motors move in short distance(1200steps for NI7358) in the shortest time. So we want to shorten the acceleration/deceleration time as possible as we can. As we test the same application before using Keyence PLC as pulse generator, it can meet our target so we expect NI also can.
During testing using MAX under 1-D interactive, when I set Velocity=3000rpm, Accel/Decel=20000rps/s, target=1200steps, I can get the best result around 1040rpm(peak velocity) with Accel/Decel time around 50ms. The time is too long for us. If I set increase Accel/Deceleration to its maximum* NI7358 will stop pulse output at all(don't know why but suspect it's because of short distance)
* Maximum Accel/Deceleration(RPS/s) = Amax X (1/Ts) X (1/Ts) X (1/R) (refer to the description of "flex_load_rpsps")
for our case: Amax=32counts/sample; Ts=0.25ms; R=4000steps/Rev, result is 128000(RPS/s)
Questions:
1. Is Amax=32counts/sample hardware fixed?
2. Any reason can cause NI7358 stop output pulse when I set maximum acceleration/decelration?
3. Does NI provide time-based profile instead of velocity one so we can set acceleration/deceleration time directly?
4. Can I use direct pulse output mode just like a pulse generator instead of current profile output?
5. Under open-loop mode, if I change PID & acceleration feedforward gain, do these will affect the output pulse performance?
6. If we change the control circuit from open-loop stepper to close-loop servo system, does it can solve our problem?
7. How does the trajectory generator produces the count/steps output during the acceleration? I create an excel file(attached below) try to draw out the relationship between the sample, velocity and acceleration, anyone can check for me it's correct or not?
My formula is here:
Position2 = Position1 + Velocity2*(Sample2-Sample1)
Velocity2 = Velocity1 + Acceleration*(Sample2-Sample1)
All comments are appreciated!
Attachments:
Profile Study.xls ‏28 KB

BO has no controls over VPN or network bamdwith, it takes what OS and DB drivers give it.
You'll need to identify at which point of the workflow to refresh the report the slowness is occuring.
You'll need to enable webi tracing and see how long the actual refresh take.
You'll need to enable network traces to see where in the network the slowdown is occuring.

BPM Concern: Total message processing time too large

Hello Experts,
I have implemented a BPM scenario with a fork step. MDM is the sender system and ECC is the final reciever system.
Branch 1 of Fork step: Processes Message type 1(DEBMAS); includes transformation step(ABAP mapping) and Send step (RFc call to ECC)
Branch 2 of Fork step: Processes Message type 2(ADRMAS); includes transformation step(ABAP mapping) and Send step (RFc call to ECC)
the messages are proccessed as required but the only concern is the total time reuired for the final message to be posted to ECC system. This scenario using the above design takes more thean 5 mins to post the message to ECC; which the end user will not accept.
Is this due to the parallel processing using fork step? Is there any alternative method of implementing this with minimum processing time?
Thanks in advance,
Elizabeth.

Hi Ninad,
Begore the ABAP mapping step, there is a send step (sends DEBMAS/ADRMAS to ECC).
ABAP mapping reads the message Id generated for the above message and the retrieves the PI IDoc number and its details from PI SXMB_MONI backend tables. This is then required to be given as input to Send step (RFC call to ECC to get ECC IDoc acknowledgment based on PI IDOC number).
~Elizabeth.

Download time too long

Why is it taking too long to download a song? Im talking HOURS if it even happens!

Could you download the graphics dynamically as they are required?
So, the air app would be the master 'app' and those additional
graphics are loaded from your server only when required? I'd bet this
would be the way you'd build it if the app were purely based online
and not via air?
Dave Cates
Managing Director
Redemption Media Ltd
www.redemptionmedia.co.uk
Interactive internet, media & communications solutions.
Sent from my iPhone

The ODI reverse process use too long time

I run the reverse process for a customized essbase model, it just keep running and never come out? should I reinstall ODI? I make another test, odiaess_93110_samples.zip, I import the ODI sample_essbase work repository. and I make a reverse process for the essbase using the next parameters,
MULTIPLE_DATA_COLUMNS = Yes
DATA_COLUMN_DIMENSION = measures
DATA_COLUMN_MEMBERS = sales,cogs
EXTRACT_ATTRIBUTION_NUMBER= No
It just keep running for a few hours, and never come out, what's wrong with me?
Edited by: user2571802 on Apr 20, 2010 7:56 PM

Are you using a defined agent instead of the local agent to reverse, I recommend using an agent as sometimes there can be issues reversing with the local agent.
Cheers
John
http://john-goodwin.blogspot.com/

Importing process takes too long!!!

I am trying to import CD's but the whole process at first goes ok, I can do a song in about 26 seconds but by the time I move on to the second CD it starts taking a long time to the point that it actually takes 18 minutes to do a song. I followed some ones advice to uncheck the box in preferences/advanced/the option to slow down the process if CD is damaged. It still does not work. I really need help because I have a lot of CD's and I want to import them!!!!

Generally, tape must be imported in real time. So a 40 minute lecture is going to take 40 minutes to import.
Export speed to a QuickTime file depends a lot on the specs of your computer and the quality settings you are using.
The the QT file is too large to put a 700 mb CD anyway.
That's where the magic of DVD encoding comes in. A 40 minute movie can easily be encoded using iDVD to fit on a DVD.
Matt

Adobe webdynpro taking time too long

Hi,
Initially, I have Read Timed Out issue. I;ve incresed the Web Container timeout to 360 seconds. but now, when I run the program with Interavtive form, it tooks very long time. It's been 30 minutes and still going. I do see the log mentioned not found ServiceContext, Any idea what's wrong?
[code]org.omg.CORBA.BAD_PARAM: Not found ServiceContext with id=0 vmcid: OMG minor code: 26 completed: Maybe
 at com.sap.engine.services.iiop.internal.giop.ClientRequest.get_reply_service_context(ClientRequest.java:284)
 at com.sap.engine.services.ts.jts.ots.PortableInterceptor.ClientJTSInterceptor.receive_reply(ClientJTSInterceptor.java:91)
 at com.sap.engine.services.iiop.internal.giop.ClientRequest.dealReceiveReply(ClientRequest.java:133)
 at com.sap.engine.services.iiop.internal.giop.ClientRequest.dealReceiveReply(ClientRequest.java:125)
 at com.sap.engine.services.iiop.server.portable.Delegate.invoke(Delegate.java:282)
 at org.omg.CORBA.portable.ObjectImpl._invoke(ObjectImpl.java:457)
 at com.adobe.service._FontManagerStub.getAdobeServerFontDirectory(_FontManagerStub.java:33)
 at com.adobe.document.XMLFormService.getFontDirectories(Unknown Source)
 at com.adobe.document.XMLFormService.getCommonDir(Unknown Source)
 at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
 at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
 at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
 at java.lang.reflect.Method.invoke(Method.java:324)
 at com.adobe.service.ProcessResource$ManagerImpl.getConfigPropertyString(ProcessResource.java:1080)
 at com.adobe.service.Manager_Tie.getConfigPropertyString(Manager_Tie.java:55)
 at com.adobe.service._ManagerImplBase._invoke(_ManagerImplBase.java:101)
 at com.sap.engine.services.iiop.internal.InvokeHandlerHolder.invoke(InvokeHandlerHolder.java:35)
 at com.sap.engine.services.iiop.internal.giop.IncomingRequest.process(IncomingRequest.java:244)
 at com.sap.engine.services.iiop.internal.giop.ConnectionParser.readMessage(ConnectionParser.java:85)
 at com.sap.engine.services.iiop.server.CommunicationLayerImpl.request(CommunicationLayerImpl.java:209)
 at com.sap.engine.core.service630.context.cluster.session.ApplicationSessionMessageListener.process(ApplicationSessionMessageListener.java:33)
 at com.sap.engine.core.cluster.impl6.session.MessageRunner.run(MessageRunner.java:41)
 at com.sap.engine.core.thread.impl3.ActionObject.run(ActionObject.java:37)
 at java.security.AccessController.doPrivileged(Native Method)
 at com.sap.engine.core.thread.impl3.SingleThread.execute(SingleThread.java:100)
 at com.sap.engine.core.thread.impl3.SingleThread.run(SingleThread.java:170)[/code]

OK. I found the problem. In one of the java parameters in the installation guide, It mentioned to add:
-Dorg.omg.PortableInterceptor.ORBInitializerClass.com.sap.engine.services.ts.jts.ots.PortableInterceptor.JTSInitializer
Well, looking at the rest of the parameter for -Dorg..Class, it always followed by '=', so I change into:
-Dorg.omg.PortableInterceptor.ORBInitializerClass=com.sap.engine.services.ts.jts.ots.PortableInterceptor.JTSInitializer
And now it is working just fine.

Sign-in time too long (Win 8.1)

Two days ago I received quite a strong new Win 8.1 computer (with SATA drive).
I have opted for a LOCAL user only without a password.
The boot process was extremely fast: the computer turns-on to the blue log-in screen in few seconds; and then pressing Enter the signing-in process was almost immediate.
Then, a bad thing happened: I was lured by Microsoft to sign-in with a Microsoft account.
Then, signing-in to my computer with this MS account (and password) took ages; I mean, something like a full one minute the computer was trying to sign-in until eventually succeeding.
Understanding my mistake (why would I want an MS account?) I decided to disconnect from my Microsoft account, and return back to my nice local user.
But now, the long waiting time to signing-in (even without a password, and with an automatic log-in) still remains: I wait like 25 sec when the blue screen and the white dot circulate repeatedly until my account is signed-in.
What is going on?
Help me!

Hi,
Did you mean after switch to local user account, it still took long time to sign in?
Firstly please disable unnecessary startup program:
1. Open Task manager;
2. Jump to Startup tab, right click to disable the unnecessary applications one by one.
If the issue still persists, refresh your computer for test:
How to refresh, reset, or restore your PC
http://windows.microsoft.com/en-IN/windows-8/restore-refresh-reset-pc
Karen Hu
TechNet Community Support

Macbook boot time too long

Hello, boot time on my MacBook seems to be very long. Can you please take a look at my bootchart and tell me what do you think?
And also the gnome login is very slow. More precisely, it takes too much time for the desktop to appear after logging in.
Here is the bootchart:
http://i49.tinypic.com/5yrk11.png
Last edited by exapplegeek (2012-07-08 22:00:21)

exapplegeek, your thread title is far too general. Please rephrase it such like: "Macbook boot time problem" as soon as possible.
And leave the "[BOOTCHART]" part out.

IPhoto launch time too long

I now have a reasonably large iPhoto library. 12409 photos, 29.8GB. It's all on the internal drive of my MacBook Pro laptop. Running iPhoto 6.0.6, OS X 10.5.5
Launching iPhoto takes a long time. It takes a long time (about 1 minute) for the iPhoto window to appear and "loading photos" to be done. Once that is done, iPhoto sits with a spinning beachball cursor for about 5 minutes before I can do anything.
Additionally, iPhoto seems to crash intermittently, proving to be slow and unreliable.
The organisation into albums and slideshows, as well as keywords/captions is very useful. Plus I have to use it to import the photos because one cannot access a Canon camera as a memory device like virtually every other camera.
Otherwise I would have deleted the 475MB application from the disk. Talk about a bloated application!
Can anyone provide an insight into how to make the load time reasonable?
Cheers,
Matt

LarryHN wrote:
Probably a user bloated disk - iPhoto works very well IF you have adequate resources - your crashing and slowness both point to free disk space
Thanks for that constructive comment. I have 10GB free disk space and 2GB ram. Activity viewer shows the CPU is mostly idle, and plenty of free memory. iPhoto seems to be hung using about 5-10% cpu time.
I do have lots of stuff on my drive, but the machine is not struggling for CPU or memory, so I hardly think that is problem.
If I "Sample Process" on iPhoto it seems that it is spinning the beach ball whilst using about 5-10% of cpu time, mostly under a CFArraySort function call. Seems to me that the program design doesn't work very well with a library of 12000+photos.
Is iPhoto 7 significantly better in this regard? If so, I'll consider upgrading in the future, but if it has the same problems then I'll look for an alternative solution. The current working order of iPhoto 6 is unacceptable.
Thanks, Matt.

Email loading times too long (Gmail, E75)

Hi dudes, and thanks for reading.
Email provider: Gmail
Phone: E75
Phone firmware: 202.12.01 (the newest)
I'm getting pretty long waiting times when my phone loads my emails.
I have set up an account to the phone's email application. It lists the emails in real time, but when I try to open an one, the loading time gets overwhelming.
It is not a pleasure to use the email application.
Can I do anything for this? Another app?
Any help wuold be appreciated!!!

Multiple reasons too hard to guess.
First step of ts - reset network settings

Windows 8.1 first boot time too long

After Sysprep, windows 8.1 tooks long time to boot, is it normal?
N.A.Malik

Hi,
Generally, when you sysprep an image and deploy it to other devices, it would prepare the device settings, registry update and other computer configuration. This phase would take some time. We call it as specialize and OOBE pass.
Here we have a metric to assess the first boot time, you can take a look at this guide:
First Boot Performance
http://msdn.microsoft.com/en-us/library/windows/hardware/jj130825.aspx
Alex Zhao
TechNet Community Support

Average processing time too long in workload monitoring (ST03N)

Similar Messages

Maybe you are looking for