What happened? As jobs are proceeding through ID, they are showing some encounters with match result errors while others in the same bulk import are showing as completed and others are still queued.
I’ve had reports from 2 users so far about errors in their bulk import match results.
Interestingly, one reported that the errors resolved “by themselves” and are now showing match results:
Here’s the bulk import link: Wildbook for Carnivores | Login
Initial errors:
So it’s hard for me to know if these errors are actual issues or just encounters still waiting for match results to be completed.
What did you expect to happen?
For each encounter in the bulk imports to show status like waiting to queue, queuing or queued if they’re still in progress. Showing an error status instead, before changing to a completed status, causes confusion and support enquiries.
Actually, most match errors should automatically re-run themselves. Some of the more recent imports were manually re-run if they appeared to have stalled. I’m not sure if the first import you linked was a result of the errors correcting themselves or if it was part of the manual batch, but I’ve manually re-run ID on the stalled errors on the second import.
I did not know that! So at what point should a user start to worry that there’s a problem that they need investigated rather than continuing to wait while an encounter with a match error re-runs itself?
I worry that these error messages, especially when mixed in with “completed” encounters in the same batch, will cause users to assume that this needs to be reported for investigation and our support volumes will go up unnecessarily if they’re mostly re-running themselves.
Is there a way to know if there’s an actual issue causing a permanent error vs when the encounter with an error is re-running itself and may still complete normally?
Sorry if I’m misunderstanding and thanks for clarifying.
If I recall correctly, it checks for errors to re-run pretty quickly. I’d say that if the other encounters are complete and the errors are still there after a few hours, it’s safe to check in with us to see if it got stuck somewhere in the process.
The biggest indicator is time lapsed. If ID is run in the afternoon and finishes with mixed results of “completed” and “error” and the user comes back to check on it the following morning and the same errors are still there, that’s something we’ll want to investigate. But if it’s a busy day in the queue and the import recently finished with ID errors, revisit it later to give the queue time to complete existing jobs and re-process failed ones.
#2 above has more than just errored match results. The completed match results pages are missing some or all of the details that should be on the match results page. Here are 2 example screenshots:
I put both imports back through identification instead of just the encounters with errors. I’ll keep an eye on them throughout the morning to make sure they progress correctly.
