Error during image creation with DALLE-3

[RankPast]

Your Task History ID

IjU3NjUwNTY5MDYzNTA0M2Q1MjY0NTUzMDUxMzc1MTY2NTQzMTBmMzMi_pc

I built a fairly straightforward automation to create images: Google Sheets (new row) > chatGPT DALLE3 > Google Drive
I used it successfully to create many dozens of images without any issues. When I run it recently, however, I've been getting errors. Some of the images come through and some don't. Sometimes it's only a couple of images that receive the error below, and sometimes it's many, many of the images that don't get created. I've tried a lot of troubleshooting but it's still happening.

Error for only some of the many prompts entered during a run of images:

Task History ID: IjU3NjUwNTY5MDYzNTA0M2Q1MjY0NTUzMDUxMzc1MTY2NTQzMTBmMzMi_pc
Step Name: OpenAI (ChatGPT, DALL-E, Whisper) > Generate Image (DALL-E)
Connection Name: OpenAI Connection
Task Error Detail:

{
"error > message": null,
"error > type": "invalid_request_error",
"error > param": null,
"error > code": null
}

 

[ArshilAhmad]

Hi @RankPast,

It seems to have been a temporary issue, as it was resolved upon re-execution. Based on your Task History, the recent executions do not appear to be encountering this problem. Please monitor your workflow for some time, and if the issue occurs again, do let us know.

 

[RankPast]

Certainly don't think it was a temporary issue as its continuing to happen - and has been for days now. Maybe the one task ID I gave you got re-executed but that's not the case for most of the errors.
It doesn't seem to happen when I only enter a couple of prompts, but if I paste in a couple dozen, as in the screenshot below, only 17 of the 24 images arrived in Google Drive.

 

[ArshilAhmad]

Got it. In that case, please add a 'Delay After Queue' as Step 2 in your workflow. This will ensure a gap of 1 minute between each task executions. Kindly add this delay to your workflow and monitor it for some time.

 

[RankPast]

Appreciate the quick response. Is this the only workaround? A major point of building the automation was that I could pump out hundreds of images in minutes....now that will take hours.

 

[ArshilAhmad]

The issue is related to OpenAI action step not being able to process multiple prompts at the same time. I'm afraid there is no other workaround except to add a delay between task executions.