Long HTTP Request/Response round trip time on new Azure Websites under BizSpark Plus

Hi,
I have a very difficult situation. I am busy migrating from a traditional VM architecture to Azure Websites. This is after a few months of testing and being happy that Azure Websites can provide the performance necessary for our application. The TEST site
I set up is running great and meets all the performance requirements. The TEST site is set up under my normal BizSpark subscription.
We have since been awarded a BizSpark Plus subscription and I have started the migration process on this subscription, as opposed to the normal BizSpark basic subscription.
However, since starting our migration process, I am experiencing huge latency on the HTTP request roundtrip from browser to server and back for my newly created LIVE Azure Website instance. I have investigated and observed the statistics between the TEST
and LIVE instances. Here is a test run against the 2 environments on a particular screen of our application. Here are the times for the full roundtrip for each AJAX HTTP request of the specific actions:
TEST (Azure Website on BizSpark subscription):
GetAudioSource - 267ms
LogStartEvent - 196ms
LogProgress - 243ms
Retrieval of Audio file - 2.06s
LIVE (Azure Website on BizSpark Plus subscription):
GetAudioSource - 938ms
LogStartEvent - 3.61s
LogProgress - 5.11s
Retrieval of Audio file - 5.97s
There is clearly something wrong with the routing of the LIVE instance starting at the load balancer.
This is not the only place affected. In fact, all the HTTP requests are very lagged and delayed in the LIVE instance as opposed to hardly noticeable in the TEST instance.
I am wondering if there is a lower priority being afforded to my BizSpark Plus subscription running my LIVE instance, of if I am just unlucky with the underlying infrastructure that has been selected on my behalf. Or, is it possible that something can be
done to enhance and optimise the routing etc?
I also need to say that I have matched the Azure Website configuration and scale settings to be exactly the same between the 2 instances and have also set Always On to ON.
Any help or thoughts will be greatly appreciated, because I am starting to seriously doubt the service levels of the Azure Websites platform.
Thank you
Ozzy

Hello Ozzy Vox,
1. Which region(data center) do you have your websites deployed in?
Ideally there would be no change in terms of type of subsciptions when you deploy your website in production.
I suggest that you set up diagnostics on to your Azure websites to get the logs to understand where it is spending most of its time and the reason behind the latency.
You can refer to this link that will give you information on 'How to Monitor Websites':
http://azure.microsoft.com/en-in/documentation/articles/web-sites-monitor/
Also refer to this link regarding 'Enable diagnostic logging for Azure Websites':
http://azure.microsoft.com/en-in/documentation/articles/web-sites-enable-diagnostic-log/
Thanks,
Syed Irfan Hussain

Similar Messages

Response times, interpretation times and round trips time

Hi
Can some body explain me what these terms means exactly ? What is their significance ? Appreciate answers without help links.
thanks
kumar

Hai,
Interpretation times: This is the duration from the time that user input actions are validated in the SAP GUI client and the moment the request is actually sent to the SAP application server. This measures the health of the SAP GUI client used for the test rather than the performance of the SAP R/3 server.
Check SAP Note: 364625.
Response Time: The response time of a transaction step is the difference in time between the point when the request arrives in the R/3 System and the point when the R/3 System completes the processing.
Check SAP Note: 8963 and 851012.
Round trip times: Is almost same as the response time. Time from client - LAN - Server and back to the client is known as Round trip time.
Regards,
Yoganand.V

HTTP request/response object not thread safe.

According to the serlvet spec. Http Request/Response
are not thread safe. Quoting from the Spec:
" Implementations of the request and response objects are not guaranteed to be thread safe. This means that they should only be used within the scope of the request handling thread. References to the request and response objects must not be given to objects executing in other threads as the resulting behavior may be nondeterministic."
This has prompt me to ask the following question.
For Example I have a servlet which does the following
request.setAttribute("myVar","Hello");
The request and response is dispatched(using RequestDispatch.include(request,response)) to another
servlet which retrieve this attribute i.e request.getAttribute("myVar");
Is this safe?
The Spec only said "The Container Provider must ensure that the dispatch of the request to a target
servlet occurs in the same thread of the same VM as the original request." I take this meaning that the targeting servlet does not have to run in the same thread(only dispatch), otherwise it would be safe.

To put it another way, you can only have onle thing working on a request at a time. For instance, the ServletContext is available to all servlets running on a server. If you tried to save a particular request to the ServletContext, it would potentially be available to many concurrently running servlets. They could all change whatever in it at the same time. Each servlet is in its own running thread. Hope that helps some.

RTT Round trip time ?

Hi team,
Have few doubts on round trip time of PING utility.
Actually am not sure how RTT is calculated. i mean min/avg/max values. is there any algorithm available ? please share if any RFC available ?
i know RTT is delay between source and destination ack. i am not sure how MIN/AVG/MAX calculated.
please help me on this.
thanks
rajesh kumar V

Disclaimer
The Author of this posting offers the information contained within this posting without consideration and with the reader's understanding that there's no implied or expressed suitability or fitness for any purpose. Information provided is for informational purposes only and should not be construed as rendering professional advice of any kind. Usage of this posting's information is solely at reader's own risk.
Liability Disclaimer
In no event shall Author be liable for any damages whatsoever (including, without limitation, damages for loss of use, data or profit) arising out of the use or inability to use the posting's information even if Author has been advised of the possibility of such damage.
Posting
RTT is figured by the sender recording when it sends the ping request and subtracts that from the time when it gets the ping reply. I.e. the delta is the RTT.
Receivers might be slow to process a ping request, so the latency isn't always just network time, it often includes the receiver's processing latency.
(NB: For Cisco's SLA "ping", the receiver immediately records when ping request is received and this usually allows the receiver's processing latency to be eliminated - providing a much more accurate network latency value.)
If you send more than one ping, in a series, the min the is the best/least time, max is the worse/highest time, and avg is the average of the series.
e.g.
If you sent 3 pings and their times were 2,2,2, min/avg/max would be 2/2/2.
If you sent 3 pings and their times were 3,1,2, min/avg/max would be 1/2/3.

Round trip time is more with wireless phone SSID

Hi,
We have separate SSID for data and voip in our wireless network.
I have connected my laptop to wi-phone SSID and gave a ping to other laptop IP on same SSID. The round trip time is around 707ms. Sometimes it is reaching more than 1000 ms.
When i connect to my data SSID, the round trip time is very very less around 3 tp 4 ms.
When I try to ping to a phone which is on call, the round trip time is around 50 ms.
Please suggest on how to resolve this issue. Please let me know, if any configuration related information is required.
Regards,
Madhan kumar G

it could be due to power-save enabled on those laptops. Disable it and try.
disable/enable wmm on laptop and try.
Try pinging the phone at off call you may experience high RTT when it is not on call.
ping on WLC is handled as best effort. try file transfer or iperf/jperf.
what's the client(s) rssi and snr when you test.
If the above didn't help check config between the WLANs in question.

Calculating Round Trip Time from Non-Spatial Log Data

Hi,
I have a log table which holds the vehicles' "entrance" an "exit" dates (dd.mm.yyyy) to/from regions like that:
VEHICLE_ID REGION IN_TIME OUT_TIME
1001 CUST_A 01.01.2007 03.01.2007
1001 CUST_B 05.01.2007 06.01.2007
1002 CUST_C 04.01.2007 06.01.2007
1001 BASE_A 08.01.2007 11.01.2007
1002 CUST_D 11.01.2007 12.01.2007
1001 CUST_A 15.01.2007 15.01.2007
1001 CUST_F 18.01.2007 19.01.2007
1001 CUST_A 19.01.2007 20.01.2007
1002 CUST_E 16.01.2007 19.01.2007
in fact, this table was created for reporting "average wait time on each customer" but then i calculated "round trip time" by assigning "base region" ( vehicle round trip starts from "base region" then comes back to here, etc.. ) and creating some rules like "if it passes from CUST_A and CUST_D, it is 'ROUND TRIP A'"..
vehicles' voyage report is like that :
TRIP STARTDATE ENDDATE
TRIP_A 01.01.2007 13.01.2007
UNDEFINED 13.01.2007 16.01.2007
TRIP_B 16.01.2007 02.01.2007
now, trips are more complicated and i want to use spatial mechanism. any advise?
thanks,
Cihan.

Stephen Rodriguez wrote:You don't need to worry about the round trip between the AP and WLC. Just make sure the phone to phone is goodHTH,
Steve
Please remember to rate useful posts, and mark questions as answered
Thanks Steve,
That makes Sense, as post authentication the phone to phone would be ofcourse less than 150ms when the traffic is locally switched.
But I dont understand why the recommendation in the D&D guide "Roundtrip latency must not exceed 300 milliseconds (ms) for data and 100 ms for voice and data between the access point and the controller"
I think it is applicable if the traffic is centrally switched?? or is there more to it, for roaming perhaps?
Thanks
Jino

Is it possible to send several http requests at the same time?

hi:
is it possible to send several http requests at the same time in j2me application, or it's device specific.
It's ok in my NOKIA SYMBIAN C++ application.
regards
Message was edited by:
danielwang

Is it possible to have 2 threads running at the same
time at different times eg 1 repeats every 20
miliseconds and the other 40 for example. Yes.
http://java.sun.com/docs/books/tutorial/essential/concurrency/index.html

Possible to get runtime info (Http request, response) from a WD4A Appl.?

Hey,
In a BSP / JSP application, it is possible to retrieve and set the runtime data (http request and response). For Example:
response.setContentType("application/x-suvaagent;charset=iso-8859-1");
response.setHeader("Cache-Control", "public");
response.getWriter().print("<Request>");
response.getWriter().print("</Request>");
Is it possible to do this in a WebDynpro (ABAP) Application?
The requirement is to construct a file using http response method. This file contains parameters like business data etc. The file type is already registered as a MIME type which is associated with a standalone program.
When sending the http response, the file will be filled out with needed data and the associated program will be called using the data in the file.
I haven't found out how to do this in web dynpro application, since we use WD4A in our project.
Thanks and Regards,
Liang

In Web Dynpro you do not have direct access to the Request or Response object. This is part of the design of the abstracted rendering and makes things like the SmartClient and the AJAX based Delta Rendering possible.
In Web Dynpro you should use the FileDownload UI element or the CL_WD_RUNTIME_SERVICES=>ATTACH_FILE_TO_RESPONSE
http://help.sap.com/saphelp_nw70ehp1/helpdata/en/47/a8d95673b12fe2e10000000a42189d/frameset.htm
If you want more control over the response object for a download then I might suggest that you can create a cached response object, place it into the ICM cache and generate a unique URL for it. You can then fire a navigation plug or a LinkToURL to go to that cached response object.

Http request response

Hi
Im having trouble with a request /response.
What I'm trying to do is send a string to a http request such
as
http://webdeploy/r/mixes/mixreq.aspx?username="+username+"&mixname="+mixname,
"_parent"
The response of this returns a string value.
Ive tried using getURL but all that does is open a new html
page with the response string printed on it.
I couldnt get this to work with urlRequest either.
Basically all I want to do is capture the returned string
thanks,
dub

Thanks for the reply Raymond.
I've been using the URIRequest class. It wasnt working for me
and then I ran the application in debug mode.
I got the following error
*** Security Sandbox Violation ***
Connection to
http://webdeploy/r/mixes/mixreq.aspx?username=bernard1&mixname=My
XYZ Mix_parent halted - not permitted from
file:///C|/Documents%20and%20Settings/ba%Fetch.swf
-- Untrusted local SWFs may not contact the Internet.
Unable to load requested document.
Do you know what this whole sandbox business is about and how
to get around it?

Certain Http Server Errors not visible in logs for Azure Websites

I noticed that a lot of times (but not every time) Http Server Errors on my Azure Website (an ASP.NET MVC app) don't show up in any logs, although they are recognized by Azure. E.g. I get alerts that Http Server Error count > 5 in the last 15 minutes,
I see the error count on the Portal under Monitoring/Requests and errors however these errors are not visible in my application's own logs or in the eventlog.xml. Thus I also can't troubleshoot them.
Now I'm not totally sure (and can't find any hints about) what exactly "Http Server Errors" mean in Azure terminology but since 4xx errors are handled separately by alerts I'd guess these are 5xx errors. If indeed then I don't understand why there
are apparently errors happening that are not visible at least in the event log?

Hi Zoltán Lehóczky,
For your issue ,I suggest that you could try to enable site diagnostics and application diagnostics in Azure Management Portal.
In site diagnostics ,it will logs Detailed Error Logging which logs detailed information for HTTP status codes that indicate a failure (status 400 or greater).
Then you could access the folder "/LogFiles/DetailedErrors" via Ftp to check Detailed Error Logs which provide any HTTP errors have occurred.
For more information ,I suggest that you could refer to the link below:
http://azure.microsoft.com/en-us/documentation/articles/web-sites-enable-diagnostic-log/
Best Regards,
Kevin Shen.

HTTP Request/Response in ABAP

Hello Experts
I have a simple requirement, where I want to access a web url from my ABAP program and get the response from url in my program. I have never handled such a requirement.
Kindly suggest the solution.
Thanks and best regards,
Anand.

Use class CL_HTTP_CLIENT and an HTTP destination in SM59. There are multiple examples of code on the forum as well as blogs/tutorials available.

Calculating Round Trip Time

Hi Everyone,
I have tried using Network Monitor and Memory Monitor in WTK but I m not sure as to how I should analyze the results or where is the time located ? I did see time taken from getting 0 to 6244(image size) bytes. is that RTT ?
Please advise.

This post has been moved from the Servers' General Discussion forum to the Solaris 10 forum.
You do not have a question specific to a hardware component of a Sun/Oracle branded server.
You seem to have a general software question that could apply regardless the brand of the computer system.
Perhaps you can review the man pages for ping, traceroute, and arp and combine them into what you hope to find. There may be yet other Solaris commands you can use, based upon what you read in their man pages.

How to read the Image Data using HTTp Request response

i want to read image data from server please send me any code or answers.
and also i want to exit application using button control for iphone simulator.
thanks in advance.

You would do a URLRequest and download the image. You could save it to disk and then load it or directly create a new image with the binary data.

Round tripping with the new Photoshop Elements 9

After installing elements 9, I am able to get photo from Aperture to Elements but after saving changes the changes do not show up on Aperture side. Has anyone had the same problem & is there a way to fix problem?

If you have correctly set up PSE 9 as your external editor in Aperture and are using it that way, then the problem lies in your save preferences in PSE. Go to Photoshop Elements>Preferences>Saving Files and set On First Save to save over the original file without showing you the Save As dialog box. That's the only way for image management systems like Aperture or iPhoto to know to create versions.

Request & Response

When we try to test the application, it sends the http request to abap engine from whcih we get the corresponding response.
Suppose that my application component has one view with two input fields and one action button.
If we run the application for the first time, we get the response in the form of a page with these two input field elements and
button. Suppose Now, i have filled tese two input fields and pressed the button and some validations happened. And closed the
web page.
Here ABAP engine responded for two things.
   i) Initially it gives the response by rendering the web page for our URL request.
   ii) Second it has done some validation for the given inputs by using the action event.
   iii) At last, I closed the webpage from the client webbrowser( here no interaction with abap engine).
    In this scenario, When can we say that HTTP request/response cycle is completed?
    Is it happened HTTP request/response for two times here?
   I appreciate ur help.

This sounds like the age old question: If a tree falls in the forest, but no one is there to hear it; did it really make any sounds? Of course the answer is: who cares?
In your description there will technically be 3 request/response cycles.
When you close the browser, some special JavaScript fires, making a request to the server to close your session and fire the WDDOEXIT method. So your assumption of no interaction with abap engine on close is incorrect.

Long HTTP Request/Response round trip time on new Azure Websites under BizSpark Plus

Similar Messages

Maybe you are looking for