locked
Major failures in accessing API today RRS feed

  • Question

  • Of late, it appears as though there have been many failures with report availability and other MSN AdCenter API failures. We presently use v4 and have run into numerous problems over the past few weeks focused around three central areas:

     

    a. Report data availability in the morning i.e. error -400063

     

    b. Certificate updates

     

    c. Unknown authentication errors (even when our access credentials are correct) i.e. error -400001

    (occurred today)

     

    Please note that while this might seem like trivial errors with easy workarounds, a lot of customers run as agencies and have a very large number of clients that they manage which makes this difficult to work with. While it is understandable that technical issues do arise, it is not easy to communicate this out to our clients when MSN AC itself doesn't acknowledge such errors or post any information regarding these errors unless requested for.

     

    Can we get an update on the problem related to the authentication errors that we've noticed today and if there's an ETA on when this will be corrected (and hopefully an explanation on what broke and what was done to correct the problem)?

     

    Thank you.

    Tuesday, February 26, 2008 10:42 PM

All replies

  • We fully understand the gravity of these issues and do not take them lightly.  Each incident is researched carefully and we attempt to take measures to minimize the chance of reoccurrences.  We also understand that the timing and breadth of our communication could improve and are currently working towards this end as well.  Please know that we are always striving to meet our customer needs and really appreciate frank and relevant feedbacks like this.

     

    To give you an update, we’re afraid that a string of unrelated incidents have contributed to creating a series of issues that you’ve been encountering.  Below us a summary of each of the problems you’ve brought up on your posting:

     

    a. The main cause of these reporting errors can be attributed to the backend data not becoming available in a timely manner.  Sometimes, our platform experiences an unusually high load and the data simply isn’t processed in time.  We are trying to address this situation and ironically, this work has also contributed to some delays because our maintenance procedures go beyond the allotted time.  Needless to say, we know that this is a pain point, and we are working hard to stabilize our infrastructure.  Also, in V5 of API, some improvements have been made to ensure that these calls do not error out regardless of the state of our data.  As such, while we promise to continue our best efforts in eliminating these reporting errors, we also urge you to give V5 a try.

     

    b. The SSL Certificate issue is being handled as a serious problem and we have dedicated resources working on this from multiple teams. A more detailed discussion can be found in another thread and we urge you to take a quick look: http://forums.microsoft.com/MSDN/ShowPost.aspx?PostID=2878812&SiteID=1

     

    c. The unknown authentication error (-40001) was actually caused by a much larger incident affecting subset of all live.com customers (including hotmail and messenger) and actually had nothing to do with adCenter API.  We understand that this issue has been fully resolved now and shouldn’t be affecting customers anymore.  As expected, appropriate teams are taking a thorough look and will take necessary measures to ensure better stability in the future.  

     

    As you can see, it was quite unfortunate that these events took place so close to each other. While this is no excuse for causing frequent outage, please know that we are doing everything in our powers to increase our stability and reliance.  

     

    Hopefully, we’ve answered your question adequately.  If not, please feel free to post a follow-up and we’ll be glad to answer them.

     

    Thank you.

     

    -Steven Hong

    Senior Technical Support Engineer [MSFT]

     

     

    Thursday, February 28, 2008 5:40 AM
  • Unfortunately we're still seeing a few issues in accessing the API for which manual intervention/retries are needed to ensure 100% retrieval of data for all accounts in our system.

     

    Here's details:

     

    a. An error message is sporadically returned when requesting for account performance data using MSN AC API v4

     

     faultCode: {http://schemas.xmlsoap.org/soap/envelope/}Server
     faultSubcode:
     faultString: Server was unable to process request. ---> Undefined Error
     faultActor:
     faultNode:
     faultDetail:
     {http://xml.apache.org/axis/}stackTrace: Server was unable to process request. ---> Undefined Error
     at org.apache.axis.message.SOAPFaultBuilder.createFault(SOAPFaultBuilder.java:222)

     ... [truncated stacktrace]

    This error is intermittent and occurs randomly for different accounts in our client base in calls made to getAPIReportStatus(...).

     

    b. In addition, we have a specialized environment that access MSN AC API v5 for a few accounts.

    This system appears to be plagued by numerous internal/intermittent errors (across all API calls) as described below:

     

    <OperationError>
           
    <Code>0</Code>

           
    <Details i:nil="true"/>

           
    <Message>An internal error has occurred.</Message>

    </OperationError>

     

    ---

     

    To summarize, we're basically stuck and can't decide on whether we should stay in production with the more stable v4 codebase or upgrade to v5 with the hope that it'll stabilize anytime soon. Also note that we cannot decouple our system to use v4 for regular API calls versus v5 for reporting API calls as is probably the case with other agencies.

     

    Are these known issues that your team has encountered before?

    If yes, what are the timelines in getting these addressed?

     

    Thank you.
    Wednesday, March 26, 2008 5:50 PM