Your data. Anywhere you go.

New Relic for iOS or Android


Download on the App Store    Android App on Google play


New Relic Insights App for iOS


Download on the App Store


Learn more

Close icon

Feature Idea: Seeing on-going, long transactions while they are active

performance
longtransactions
diagnostics
load
feature-idea

#1

Hello folks,

I’ve noticed an artifact that is kind of ‘by design’ so please just listen and hold my hand while I complain for a bit, ok? :slight_smile:

I found a query today that someone was able to issue via a test UI that between our DB and S3 was actually able to take 4000 seconds to complete during a load cycle.

“Outrageous!”, you say. I agree. “Idiotic!”, you say. I agree.

Now, to the fun part - NewRelic could show me the load on my server and it could show me a ‘blip’ on transaction load when the transaction completed, but while the operation was in progress (even for over an hour), NewRelic never told me anything about the request that it new was in progress, but that had never completed.

“By design!”, you say. I agree.

However, is this something that could be made available? Is it possible to configure the system so that it will log requests that take more than 20-30 seconds to complete as a partial log so that you can at least see what is long-running and active on the machine at the moment? All of the .NET information about where the request started would be available, but it might not have finished yet. After the request finished, the record could be overwritten.

“Ridiculous!”, you say. I agree.

But could seeing ‘active’ load for long-running or stalled/problem transactions be useful? It definitely would be when scenarios come up where load is seen, but the source of the load won’t be seen for a reasonably long period of time.

“Silliness - do it another way by cleaning these transactions up more quickly!”, you say. I agree. “Don’t let your transactions run this long - how could that be useful?”, you say. I agree.

In the midst of all of this, is there a helpful space here for identifying ‘ludicrous transactions’ that are taking longer than a set threshold and are still active? If so, I would be must appreciative without having to use a debugger or second-guess from looking at actively-running sprocs/queries.

“Enough of these trifles - finish your post already!”, you say. I agree :smile:


New Relic edit

  • I want this, too
  • I have more info to share (reply below)
  • I have a solution for this

0 voters

We take feature ideas seriously and our product managers review every one when plotting their roadmaps. However, there is no guarantee this feature will be implemented. This post ensures the idea is put on the table and discussed though. So please vote and share your extra details with our team.


#2

Oh @Jeremy_Buch - I love your feature requests! I am going to send this on, 100% verbatim, to our product managers. Thank you for the great story and the smiles!


#3

One more vote for this feature request :slight_smile:


#4

@jure.mijic - I have your +1 and will get it before our product managers. :slight_smile:


#5

Hi All,

Do you have any update on this topic?


#6

No update to share right now, @apaklonski! Please be sure to vote in the poll above! :blush:


#7

I agree that this would be a very useful feature. While yes you could argue that in a web app you shouldn’t have long running processes, if NR is trying to promote its ability to monitor non web transactions including long running background processes then I think this is a must have. It also allows for Alerting when long running processes exceed a threshhold while its running (beyond some default measure).


#8

Noted, @shaney! Thank you for voting and for adding your use case to this thread! :blush: