Describe the problem/error/question
I have a workflow that requests DALL-E 3 images and retrieves the DALL-E 3 images. I regularly have issues where these don’t work, much more often than should be the case; it probably fails 30% of the time. When I rerun the entire workflow, it often works. What’s up with that? Is this a normal occurrence?
What is the error message (if any)?
The error is usually “Bad Request: Please check your parameters” but sometimes I get other errors. But one thing is common - if I rerun the workflow enough times, it works. Which to me suggests a problem somewhere on n8n, on OpeAI or somewhere in between.
Please share your workflow
Multiple workflows
(Select the nodes on your canvas and use the keyboard shortcuts CMD+C/CTRL+C and CMD+V/CTRL+V to copy and paste the workflow.)
Share the output returned by the last node
Information on your n8n setup
- n8n version:
- Database (default: SQLite):
- n8n EXECUTIONS_PROCESS setting (default: own, main):
- Running n8n via (Docker, npm, n8n cloud, desktop app):
- Operating system:
Interesting.
Are you on cloud n8n or docker hosted n8n? If docker is it behind a domain?
Could you copy and paste your workflow into the field that appears when clicking ‘</>’ on the forums here to see the logic that is occurring. And potentially screenshots of detailed errors.
I am using hosted n8n. Here is the workflow and here is the other error I get regularly, “The connection cannot be established, this usually occurs due to an incorrect host (domain) value”.
Just curious, are you a bot or a real person?
I’m real (name is Justin), I am mostly busy, but try to find time each day to help out on the forums here. It was a coincidence I replied to both of your posts lol. But this is the official account of ThinkBot.
That helps, but could you provide the actual flow? I am curious if there are dynamic fields or inputs and that may cause some downstream issues.
GPT also has been under heavy load since their image and new chat releases. I might wager if its random (with intermittent success), that you could be accessing a high load route or based on locale. Try adding a retry on error or other error handling.
did you find a solution? I found a work around where I just use a error output on the http request then loop it around. but, that can make hundreds of tries before I get the 8 images downloaded i need downloaded.
NM I found the answer here I think. no more http requests to get this done. DALL-E 3 integration
I just added the openAI node and can request an image from that.
