Community Spring Cleaning week is here! Join your fellow Maveryx in digging through your old posts and marking comments on them as solved. Learn more here!

Alteryx Server Discussions

Find answers, ask questions, and share expertise about Alteryx Server.

Alteryx Server Usage Report showing 0 execution times for all jobs

Paul_Holden
9 - Comet

Hi,

 

I'm using the standard Alteryx Server Usage Report (ASUR) (version 2018.3) running on our gallery to take a weekly look at the execution times of workflows on our server.

 

As of the last couple of weeks the report is showing ALL execution times as zero. This is in both fields - ExecutionTime and Result_ExecutionTime.

This started with a workflow run of 9th July following a number of workflows which showed a large negative execution time (arithmetic overflow possibly?). All previous records show a non-zero execution time, all subsequent records the execution times are zero.

 

I have tried working on a simple workflow extraction from the MongoDB (since the relevant data is held in a json blob it's not possible to interrogate the MongoDB directly for these values) using the ServiceDataParser macro from the ASUR. I get the same results as from the full Tableau extract i.e. 0 execution times for all jobs starting from 9th July. So this is not an artifact of the later processing of the data, although it might be down to an error in the macro?

 

3 REPLIES 3
Paul_Holden
9 - Comet

So, not exactly a full Solution but we have managed to resolve this issue.

 

Today we stopped / started the AlteryxService service on the Gallery server in order to take an ad-hoc backup of the MongoDB.

 

Following this the system is once again recording [ExecutionTime] and [Result_ExecutionTime] values for workflows.

 

Note that records that previously showed 0 values are still doing so, so that this definitely appears to be a bug in the Alteryx Service where it does not record the ExecutionTime value correctly under some unknown conditions.

 

So the good news is that restarting the service appears to resolve the ongoing problem.

The bad news is that the historical ExecutionTime values are lost and the closest you can get seems to be to take a DATEDIFF of [Result_ExecutionStartTime] and [Job_CompletionDateTime]

 

Not sure if it is valid to mark this as an accepted solution. Happy to take any mod advice on that.

Testing on records with valid ExecutionTime values shows this to be "good enough" if you want times to the nearest 10 seconds or so, which I would think would fit all likely use cases.

 

Balders
11 - Bolide

We've hit this in Alteryx 2020.4 

 

The issue has been running for a few days now and an Alteryx stop/backup/start hasn't resolved the issue. Any other pointers would be appreciated or we'll be moving our reports to use date diffs. 

Paul_Holden
9 - Comet

We have had this happen several times since I first posted. On each occasion it has been resolved by restarting the Atleryx service on the Gallery host. Note that we do not have any additional Workers, just one server acting as Gallery host and Worker.

 

I can't think of anything else to suggest I'm afraid.