Upload to API (Airtable) takes longer than 60 minutes and fails

Ask a question
1

Is this an airtable limitation or a parabola limitation?

Below is there error I received. Any thoughts on how to fix?

Screen Shot 2021-01-19 at 12.33.30 PM

2

Hey @Henry_Johnson,

This usually happens when you run into a rate-limiting issue within Airtable. They allow no more than 5 requests per second or 60 requests per minute.

Try dropping down your Max requests per minute value under the step’s Rate Limiting section to something like 60 to see if that helps.

If not, we may need to come up with a more creative way to merge multiple requests into a single row!

3

Hey Daniel,

I have it set at 300 per minute since its 5 per second (60*5 = 300). Am I thinking about this wrong?

Where did you get 60 per minute?

Also, I am not receiving a 429 error which I would expect to get if I hit their limit. Could it be something else going wrong? It seems that way to me.

4

Hey Henry

Good catch on that, and thanks for pointing it out! That calculation was definitely my oversight. :slight_smile:

With that in mind, you likely did not receive a 429 error since you were still within the rate limit, but the number of rows may that were being processed within the rate limit may have taken longer than we allow.

Parabola sets a 60 minute on how long a step can calculate before we decide to err out. If you were processing tens of thousands of rows at 300 requests per minute, that may be cause some issues.

This may also be a transient issue, so try running your flow once more to see if you can sucessfully make your API request.

5

Ok, I will try again. In case it doesn’t work, ss there any way to get rid of that limit?

6

Finessing this is always a balancing act, but there are a few workarounds you can try.

  1. Make sure that no other flows or applications are making a request to your base at the same time. If you are syncing other data to your base outside of Parabola, that counts towards your requests per minute.

  2. Break up your requests into separate API calls. I like to use the Insert rows and Filter rows steps together to send a limited amount of rows. You can then turn your API requests into multiple branches with less data.

  3. If that doesn’t work, it may be best to split up your requests into separate flows and run them on an offset schedule.

Let me know if this helps!