Node Max Tries / Wait Values Limits Too Low

Bug Description

Within the Airtable node, it seems I can’t set the Max. Tries value to anything above 5 and the Wait Between Tries (ms) value to anything above 5000. This is frustrating, because sometimes I need to set much longer retries due to Airtable API issues. However, that’s not possible in the n8n UI:

image

I suspect these limits were set at some sort of global level (applied equally across all node types), but I’m not sure.

Rather than trying to come up with better upper limits, could we just have the upper limit check code disabled through some sort of global configuration setting (like an ENV variable)? That would help a ton.

To Reproduce

  1. Create a new workflow
  2. Specify an Airtable Create or Update node
  3. Click gear icon
  4. Enable Retry On Fail
  5. Specify a Max. Tries value of 20
  6. Specify a Wait Between Tries (ms) value of 20000
  7. Save the workflow
  8. Open back up the Airtable node
  9. Click the gear icon
  10. See that both values are set much lower than specified.

Expected behavior

I expect n8n to allow me to set arbitrarily high values for Max. Tries and Wait Between Tries (ms). I really don’t want any annoying guard rails here.

Operating System

Google Cloud Linux

n8n Version

1.41.0

Node.js Version

18.16

Database

PostgreSQL

Execution mode

main (default)

Original Reference:

I had the same problem the other day with Email send node. For some reason the smtp server stuck and the Max. Tries 5 - Wait 5000 was not enough.

I think there must be an .env variable to define these limits if someone wants to increase them.

1 Like

Can we get more infos on this ?

Also running into issues with these low limits.

Google Sheets often has downtimes of 2-3 minutes which makes workflow planning much harder without being able to have the execution handle it directly.

The 5000ms retry is the same it seems across all nodes. Have checked a good number and all default to max of 5000ms

As a workaround of sorts, you can change the error processing of the node to use an error output, add a wait node which you can set to anything and then loop it back to the node input.

Would this work?

Although you wont have your max # of tries using the wait loop method. you could add a an if node in there to check against a counter if you needed to keep that.

2 Likes

Hello,

I’m also having this issue…Very inconvinient :frowning:
Hope @system can fix this ?
Thnaks :smiley:

Indeed, those limits does not make any sense, especially for self-hosted.

Some API are pretty strict in rate limiting, and the Agents fail because of that, If I could raise the wait time I could workaround the issue and my workflow would be working.
Currently, my only solution would be to throw more money at the API provider. Does not make sense.

Interestingly enough, there is a note in n8n source code to remove the hardcoded limits.

Is there a target date to fix this?