I'm not having as many problems with this issue as tozor, but I have seen reports sporadically taking abnormally long to complete recently. I run a set of reports every morning to get updated performance data, and usually they complete within a couple minutes. Yesterday I had a report that took over 2 hours, and this morning it took about 30 minutes. Just wanted to throw that out there so others know that it is not necessarily an isolated issue.
On Aug 28, 8:30 pm, tozor <[EMAIL PROTECTED]> wrote: > So I have one other interesting piece of data. Today I let the > "stuck" job continue to run in hopes that you folks could do some live > debug on your end. The report did finally complete after 9.5 hours. > This is the same report that typically takes one or two minutes. > > Hope this helps your report team isolate the issue on your end. > > Tim > > On Aug 28, 11:11 am, tozor <[EMAIL PROTECTED]> wrote: > > > I appreciate the reply but am frustrated because I don't see from your > > answer how I am going to get this resolved. I currently have a report > > which is again stuck. Can you at least tell me from your logs when > > the Adwords system started generating the report and when the system > > completed generating the report? This is the only means I have to > > prove whether or not the problem is on your end. The report id is > > 433141491. It seems like if I just let this run that someone on your > > report team could shed light on the following possibilities: > > > 1) The job completed on the Adwords end but is still returning IN > > PROGRESS > > 2) The job is indeed IN PROGRESS on the Adwords end but is never > > getting a chance to run > > 3) The job failed on the Adwords end but is still returning IN > > PROGRESS > > > I could single thread my application in this area, but my application > > manages 20 accounts which are synced multiple times a day. So this is > > not a great solution for me. I can also build in logic to detect > > these stuck reports and it seems like I will probably have to do this, > > but it would be nice to get an engineer on your end to help me > > understand which of the above three possibilities is happening for the > > specific report which is currently stuck. > > > Tim > > > On Aug 27, 12:46 pm, AdWords API Advisor > > > <[EMAIL PROTECTED]> wrote: > > > Hello Tim, > > > > I unfortunately don't have much by way of new information from the > > > reports team. Looking into the logs associated with the report job ids > > > you've specified hasn't given any indication that there's something > > > abnormal preventing those reports from completing, and they say there > > > were not any changes around Aug. 11 that would account for different > > > behavior. Based on your descriptions of what's going on, I find it > > > hard to believe that there isn't something amiss, but they haven't > > > been able to turn anything up. > > > > The advice I got back doesn't address why you suddenly would have > > > started seeing different behavior, and is along the lines of > > > scheduling smaller batches of concurrent zero-impression keyword > > > reports. While there is not a hard limit on the number you can run, > > > they do require much more processing than other reports to complete. > > > > Cheers, > > > -Jeff Posnick, AdWords API Team > > > > On Aug 26, 5:25 pm, tozor <[EMAIL PROTECTED]> wrote: > > > > > I have a stuck one I'll let go for awhile. The report id is > > > > 431736501. It has been running for 2.5 hours. Hopefully your report > > > > team can find out what is going on with this one. > > > > > Thanks. > > > > > Tim > > > > > On Aug 23, 8:14 am, tozor <[EMAIL PROTECTED]> wrote: > > > > > > One more example that occurred this morning. This time I only had > > > > > three concurrent reports. The report ran for 55 minutes before I > > > > > killed it. The stuck report id was 428986461. After I killed the > > > > > process the next report finished in two minutes. > > > > > > I hope I can get some resolution to this next week. Baby sitting my > > > > > application over the weekend is getting old. > > > > > > Thanks for your help. > > > > > > Tim > > > > > > On Aug 21, 3:09 pm, AdWords API Advisor <[EMAIL PROTECTED]> > > > > > wrote: > > > > > > > Thanks, I got pinged back from the reports team asking if there was > > > > > > a > > > > > > more recent example. I'll provide that report id to them. > > > > > > > Cheers, > > > > > > -Jeff Posnick, AdWords API Team > > > > > > > On Aug 21, 3:45 pm, tozor <[EMAIL PROTECTED]> wrote: > > > > > > > > Jeff, > > > > > > > > Just letting you know that the problem does still exist, although > > > > > > > I > > > > > > > went for a couple of days without seeing it. The latest stuck > > > > > > > report > > > > > > > id is 474410941. However I had to kill the process to keep my > > > > > > > application users happy. The report completed on the subsequent > > > > > > > attempt within a minute (the stuck report ran for about 40 minutes > > > > > > > before I had to kill it). > > > > > > > > Thanks again for looking into this. > > > > > > > > Tim > > > > > > > > On Aug 19, 12:00 pm, tozor <[EMAIL PROTECTED]> wrote: > > > > > > > > > I saw the problem on Sunday night and was going to wait until > > > > > > > > Monday > > > > > > > > morning to post with an updated report id. I have not seen the > > > > > > > > problem in the last two days. However I also backed down the > > > > > > > > concurrency to three concurrent reports in one part of my > > > > > > > > application > > > > > > > > where it is super critical for me not to have a stuck report. > > > > > > > > There > > > > > > > > is another part of my application where I left the concurrency > > > > > > > > at 5 > > > > > > > > reports (both of these are down from 10). > > > > > > > > > I will re-post on the next re-occurrence to let you know if the > > > > > > > > problem still exists. Thanks for checking in. > > > > > > > > > Tim > > > > > > > > > On Aug 19, 11:51 am, AdWords API Advisor > > > > > > > > > <[EMAIL PROTECTED]> wrote: > > > > > > > > > Hello Tim, > > > > > > > > > > I'm sorry that there's no significant update from the > > > > > > > > > reports team > > > > > > > > > yet, but I did want to check in to see if the behavior has > > > > > > > > > changed at > > > > > > > > > all since last week. I can understand the need to manually > > > > > > > > > cancel the > > > > > > > > > offending reports to ensure that your clients get the retried > > > > > > > > > reports > > > > > > > > > in a timely manner, so in the interim you shouldn't hesitate > > > > > > > > > to cancel > > > > > > > > > them and retry. If the reports team comes back to me saying > > > > > > > > > that they > > > > > > > > > need a "live" example I'll sync up with you about that. > > > > > > > > > > Cheers, > > > > > > > > > -Jeff Posnick, AdWords API Team > > > > > > > > > > On Aug 15, 6:00 pm, tozor <[EMAIL PROTECTED]> wrote: > > > > > > > > > > > Again thanks for the reply. One complicating factor is > > > > > > > > > > that I can't > > > > > > > > > > let a job remain stuck indefinitely because it backs up my > > > > > > > > > > application. At most I have about four hours after a post > > > > > > > > > > that I can > > > > > > > > > > wait. If you folks were in the middle of debugging it, > > > > > > > > > > then I can let > > > > > > > > > > it stay stuck, otherwise in the interest of my application > > > > > > > > > > users I am > > > > > > > > > > forced to kill the process and reissue the report. As far > > > > > > > > > > as the > > > > > > > > > > report ids you mentioned: > > > > > > > > > > > 421168971 - this completed after about 50 minutes, again > > > > > > > > > > this is > > > > > > > > > > highly abnormal since it usually takes 1 minute > > > > > > > > > > 471606181 - I had to kill this one after 2.5 hours. > > > > > > > > > > 471187411 - I had to kill this one after 90 minutes > > > > > > > > > > 471986881 - This was the one I reported today and I let it > > > > > > > > > > go for > > > > > > > > > > almost 4 hours before I had to kill it. > > > > > > > > > > > I will say that the accounts in question are very large and > > > > > > > > > > this may > > > > > > > > > > be a contributing factor. From my black box perspective it > > > > > > > > > > seems like > > > > > > > > > > you folks have code which predicts the cost of reports and > > > > > > > > > > doesn't let > > > > > > > > > > the resource cost from a single user exceed a certain > > > > > > > > > > limit. It seems > > > > > > > > > > like the jobs which get queued up as a result of the > > > > > > > > > > resource cost > > > > > > > > > > never get started. I am thinking that these jobs get > > > > > > > > > > cleared from the > > > > > > > > > > queue when the session goes away when I kill the process. > > > > > > > > > > Clearly I > > > > > > > > > > am guessing here but this is the type of behavior I see on > > > > > > > > > > my end. > > > > > > > > > > > I will continue to post stuck job ids to this thread if > > > > > > > > > > that is useful > > > > > > > > > > for you. If not let me know so that I don't spam the group > > > > > > > > > > too much. > > > > > > > > > > > Tim > > > > > > > > > > > On Aug 15, 4:33 pm, AdWords API Advisor <[EMAIL PROTECTED]> > > > > > > > > > > wrote: > > > > > > > > > > > > Hello Tim, > > > > > > > > > > > > 421168971 completing successfully should not have been > > > > > > > > > > > the result of > > > > > > > > > > > any special intervention on our end, no. How long did it > > > > > > > > > > > take before > > > > > > > > > > > it finally was marked as complete? And what about > > > > > > > > > > > 471606181? > > > > > > > > > > > > Okay, thank you for the debugging that you're doing on > > > > > > > > > > > your end. I've > > > > > > > > > > > only heard back about 471187411, and that job did not > > > > > > > > > > > complete > > > > > > > > > > > properly on our end, so it's not a question of the job > > > > > > > > > > > completing but > > > > > > > > > > > the wrong status being returned back. I have not heard > > > > > > > > > > > anything about > > > > > > > > > > > what the potential cause of the problem is yet > > > > > > > > > > > unfortunately, so I > > > > > > > > > > > can't provide any concrete recommendations as to how to > > > > > > > > > > > work around it > > > > > > > > > > > at this time. I will update the open bug with the latest > > > > > > > > > > > job id > > > > > > > > > > > though. > > > > > > > > > > > > Cheers, > > > > > > > > > > > -Jeff Posnick, AdWords API Team > > > > > > > > > > > > On Aug 15, 12:28 pm, tozor <[EMAIL PROTECTED]> wrote: > > > > > > > > > > > > > Sorry to be a pain about this, but this is becoming a > > > > > > > > > > > > real issue with > > > > > > > > > > > > my application. I have been successfully running > > > > > > > > > > > > multiple concurrent > > > > > > > > > > > > keyword status reports since V10 and all the way thru > > > > > > > > > > > > V12. Now > > > > > > > > > > > > seemingly without reason (at least as far as my > > > > > > > > > > > > existing application > > > > > > > > > > > > code goes) I am getting these situations where the > > > > > > > > > > > > report stays in > > > > > > > > > > > > progress forever. I have tried reducing the number of > > > > > > > > > > > > concurrent jobs > > > > > > > > > > > > and even > > ... > > read more ยป --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "AdWords API Forum" group. To post to this group, send email to adwords-api@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/adwords-api?hl=en -~----------~----~----~----~------~----~------~--~---
