Question

This ChatGPT step hit an error - unable to resolve

  • 6 December 2023
  • 3 replies
  • 169 views

Hi all

I’m facing this error when using ChatGPT.

 

Error message: The app did not respond in-time. It may or may not have completed successfully.

 

Note: This particular error started appearing since the last few days. And at every point the error occurred, I checked OpenAI Status and it showed all their systems are operational.

 

I have a paid OpenAI plan, API access. 

 

It appears to me that Zapier is not giving enough time to get the response from OpenAI and rushing through to the next step.

 

Here is my workflow, (I’m using the latest ChatGPT version on Zapier): 

 

I tried contacting support, they are super slow and their feedback is not being helpful. 

 

 

For the past few days, it’s just being impossible to get any of the Zaps using ChatGPT to work. We’ve burned through 2000 tasks with zero output as Zapier will just keep failing to retrieve a response from ChatGPT as it just won’t wait for the response. 

 

Initially I thought perhaps OpenAI is under stress due to high traffic, but, day after day, the same issue keeps recurring. We try to work through our Zaps, we use up our tasks and our OpenAI API credits, only to have the Zap fail again and again. 

 

Thank you!


3 replies

Userlevel 7
Badge +6

Hi there @Aubrium,

Welcome to the Community! 🎉

The error you’re seeing looks like a timeout error. Additionally, it looks like ChatGPT has a recent incident that might be related to your issue, which now has been resolved:

https://status.openai.com/incidents/bq8y55s5kg0l

Are you still seeing this error on your end? Please let me know!

Hi there @Aubrium,

Welcome to the Community! 🎉

The error you’re seeing looks like a timeout error. Additionally, it looks like ChatGPT has a recent incident that might be related to your issue, which now has been resolved:

https://status.openai.com/incidents/bq8y55s5kg0l

Are you still seeing this error on your end? Please let me know!

Hi there,

 

That is not the GPT version that I’m using though. The version I’m using didn’t have any outages in the past few days. 

 

The issues seems to be that Zapier won’t wait enough for GPT to respond to the query and push through the next steps and error out.

Userlevel 7
Badge +6

Hi @Aubrium,

Thanks for letting me know.

It looks like you’ve reached out to our Support Team about this and they replied on Thursday with a suggestion. I’ll share details of the suggestion here in case it’s helpful:

I had a peek at the failed Zap runs and I can see this is likely happening because the Zap is triggering on multiple updates, as the Notion trigger is a polling trigger.

To see if we can resolve this, can I please ask you to add a Delay After Queue in Delay by Zapier step right after the trigger step? You'll then want to publish the Zap for the changes to take affect.

For the workflow itself, is there a specific reason the Assistant API is needed for the text completion? For instance, are you creating custom agents with specific data? If so, by adding a lot of data to a specific Assistant, the initial step may time-out and then subsequent steps may experience a different error.

If you’ve got any questions regarding their suggestion it would be best to continue the conversation with them over email.

Thank you for your patience and understanding.

Reply