Skip to main content
Archived

Job time forecasting

Related products:Integrations
  • October 29, 2017
  • 3 replies
  • 55 views

chriswilson
Enthusiast
Forum|alt.badge.img+22

This morning a colleague indicated to me that it would be useful if we could indicate how long a job is likely to run for on FME Server. Clearly there are a lot of moving parts including source data sizes, transformers used (blocking vs feature), number of transformers, readers/writers etc. going on which could make it hard to auto-forecast, but I wanted to throw the idea out there.

Of course a simpler solution is a private parameter that the author of the workspace supplies, to give an indication once they have run the workspace enough times to know how long it should take.

If this was done it could appear on the workspace summary window after submission, and/or be emailed to the user who wanted to know.

This post is closed to further activity.
It may be an old question, an answered question, an implemented idea, or a notification-only post.
Please check post dates before relying on any information in a question or answer.
For follow-up or related questions, please post a new question or idea.
If there is a genuine update to be made, please contact us and request that the post is reopened.

3 replies

david_r
Celebrity
  • October 30, 2017

If you activate the FME Server Dashboards there's one called "Average Run Time Per Workspace" that might be of interest. It shows the average run time of the 20 longest running workspaces on your server over the last two weeks.

An alternative solution might be to make a custom dashboard workspace that aggregates more detailed execution times for your workspaces and presents a graph or table with some statistics, e.g. mean and min/max durations.

Should be fairly easy to implement, look under the "Creating Your Own Dashboard Reports" heading here for some pointers as well as more info about the FME Server Dashboards:

https://docs.safe.com/fme/html/FME_Server_Documentation/Content/WebUI/Dashboards.htm

If you decide to make your own dashboard you could base it on the "AverageRunningTime.fmw" workspace to get a head start.


rylanatsafe
Safer
Forum|alt.badge.img+14
  • Safer
  • October 30, 2017
Great suggestion, @david_r! I like this approach as it gathers "real-world" data rather than providing an estimate (in advance of the run). Unless the machines hosting FME Server and FME Desktop are identical, there is always chance the processing numbers will vary.

 


chriswilson
Enthusiast
Forum|alt.badge.img+22
  • Author
  • Enthusiast
  • October 30, 2017
Thanks David and Rylan for comments. That could be a good way of going about it, so I'll take a look at that.

 

 

Otherwise I'm thinking of times where the immediate situation at runtime varies from previous runs, e.g. available RAM is somehow reduced or whatever else has changed and the new run might be significantly longer than others - in which case an alert could even be created. I'm sure it could be made clear that it's an estimate.