Workflow succeeds, but encoder node fails - how to "report" that to Web interface?

[gdpodesta]

My job is "flowing" as expected, handling errors as it should, but the web interface is reporting it as a success?
This all seems to be as it should be, but how can I change the Workflow such that if an error condition occurs (such as a failed encoder), that it is reflected in the Web interface?

2-21-2020 6-12-36 AM.jpg (65.4 KiB) Viewed 14906 times

Maybe I'm confused about what to expect from the Web Interace. Looking at his, it appears that all the jobs finished successfully, though the middle two encountered errors, which were handled, but no transcoding occurred, and the top job is actually still in progress.

2-21-2020 6-28-39 AM.jpg (75.39 KiB) Viewed 14905 times

[momocampo]

Hello Gdpodesta,

Hehe, it's a question of interpretation. Yes, from the point of view of FFAStrans, even your workflow arrives on your "invalid media"node, it IS a success because the workflow goes to the last node without error(FFAStrans error). So all is ok for FFAStrans so all is ok for the webinterface
Usually in your case we put a email node instead of "invalid media"node. This email warns there was a problem.(encode problem in your case).
Maybe Encodem could do something for you or Steinar will be able to explain better than me but for me all makes sense
Hope it helps.

Cheers.

Benjamin

[gdpodesta]

It makes perfect sense to me from an FFAStrans perspective, but from an operational standpoint, it's a nightmare.
We get SO many status emails, that it would be lost in the noise.
If I could make two suggestions to the interface:

For each line in the "Finished" section, indicate what step/node that the line is for. Having 3 lines (in my case), all showing the same workflow and file is not very helpful - I'm stll not sure what is triggering a line to appear there.

2 Add a flag to "appropriate" nodes that would flag the Workflow as "Unsuccessful" and report that on the interface. In my case, and as you pointed out, the workflow DID complete successfully, Or, another processor, (like Send Email) that accomplishes the same thing.

All due respect, it doesn't help to know the Workflow is happily finishing, if the results that are expected aren't there, and I have to go check email to see if something went wrong, or what it was that might have.

[momocampo]

Hey,

Ok I give you a solution:
AFter your "invalid..."node, populated a variable and
%s_success% ->>>> Error (or any message you want)

In that way, you will have the message that you want inside the webinterface



B.

PS: Maybe Steinar has another solution to altered the final state but I don't have any ideas for the moment....

[gdpodesta]

PERFECT! That's exactly what I was looking for Thank you good sire.

[momocampo]

Coooolll,

You're welcome

[crispyjones]

I have a similar issue that I think would be best answered here. I have slightly different problem with the catch 22 of successful completion of error handling blocks. When my workflow encounters an error it moves to a Populate Variable block via execute on error, that block sets the %s_success% variable to equal %s_test_variable_1% error on %s_prev_node% block is %s_error% (test variable is populated with the windows user associated with the file) and it reports everything as expected. Unfortunately the simple WebUI at 65445/monitor that line shows green. This is confusing for my users as it shows up as green and they think it is just another successful job because it isn't red. Is there anyway to have the line show up as red?

error.png (38.59 KiB) Viewed 14541 times

[emcodem]

Hey @crispyjones
your request is absolutely valid as the workaround proposed above is just a cosmetic thing, it sets a message but it does not set/keep the error state or the job.

unfortunately this behaviour is intended as well and there is a feature missing for the stuff that you do. It is pretty new the we introduce users to use the populate variables processor to set the job message but there is no way to set the error message yet.
The reason is that in your workflow, you "catch and resolve" the error by just putting a node on the error path. As this node succeeds, the workflow also reports success because the error was caught by you.

Now as you seem to be a pretty smart guy i would like to introduce you to one of the most powerful new features of 1.0 release: Custom Processors.
Lets start easy: i built a custom processor for you, you extract it to: \Processors\plugin_nodes\custom_nodes\
So after extracting this folder exists: \Processors\plugin_nodes\custom_nodes\Force Error

Then you restart the workflow building ffastrans interface and insert a new processor from "Others" -> Force Error
Open the Processor GUI once and save it.

This processor is pretty simple, it forces the workflow into error and sets the message you input on the gui. But do not catch the error path after this processor again, as it would again leat the workflow to success.

Force Error.zip

(2.93 KiB) Downloaded 589 times

Note that we did not yet introduce a lot of this Custom proc feature because
1) We did develop the feature and test it but did not come up with useful processors yet.
2) We are not 100% certain if everything is bug-free, especially when having multiple custom processors installed. One should not make any problems.

So currently, please be a little bit cautious (testing etc...) when using custom processors.
I'll work on the corresponding wiki page which allows you to add your own processors easily in future

[crispyjones]

This is great, thank you. Not only does it deal with web ui issue, I now have custom processor I can pull apart and look at. Thanks again!

[emcodem]

Hehe i knew you would look into the custom processor stuff... I updated the wiki and added lots of information plus another example processor.
If you walk trough it at some point in time, it would be cool if you let me know if something is missing or if there is any issue or so...

Just know that you should not think about building processors that do transcode yet because the custom processors are missing all access to the central functions that calculate the audio mapping etc...