Web Requests Gone Missing

Hi!

We are struggling a bit with web requests. A bit too often the action node throws an error saying that the request was sent, but no response was received. The request is sent to a Cloud Run service on Google Cloud Platform, there are no traces of the request in the logs on this service.

Does anyone have any tips on how to fix this?

Hi @martinbarthel ,

Thank you for sharing your situation! While I’m not specifically experienced with Cloud Run, I’d suggest a few strategies to help pinpoint the issue:

  • Confirm that your Cloud Run service’s expected headers and HTTP methods are correctly implemented. Incorrect or missing headers often cause requests to be declined before reaching your application.
  • Look into the timeout settings to ensure they’re not too brief, leading to premature termination of requests. This is particularly vital if your service experiences longer startup times from a cold start.
  • Use Postman to send requests, which can help determine if the problem is unique to your current setup in Create.
  • If Cloud Run allows, enable more detailed logging to obtain clearer insights into operational flows and identify potential bottlenecks or errors.

I hope this helps you resolve your issue!

Hi @Simona !

I have looked more into it and it does not look like the problem is because of Google Cloud Platform but rather the “Generate Document”-action on Appfarm. This is the action that after some debugging throws the ApiRequestError. Here is a screenshot of the console:

Hi @martinbarthel

I found the error message in our documentation. Could you try to add .appfarm.app to the list of accepted domains in your Google Cloud Platform and check if that helps?

I am not sure why that will help. The Generate Document action node seems to reject the requests, and the web requests to our application on GCP seem to have been executed successfully. It looks like this because it is the https://pdf-service.appfarm.io/api/v1/generate that has been blocked, and this action node would not have been reached if an exception was thrown when requesting our GCP app.

Could you explain why adding .appfarm.app on our GCP app would help and what’s wrong with my understanding?

Hi @martinbarthel !

We are inspecting whether this is related to the scalability of the PDF service - can you provide us with a bit of info on

  • Does this happen every time?
  • Do you have any timestamps when this has happened recently (for us to inspect the logs)?

-kristian

Hi @kristian !

This does not happen every time, it usually happens for the first users trying to generate a PDF on a given day.

I have some timestamps (DD.MM.YYYY HH:MM):

  • 08.04.2024 08:00
  • 03.04.2024 08:57
  • 26.03.2024 15:21
  • 26.03.2024 15:19
  • 26.03.2024 05:34
  • 21.03.2024 11:05
  • 14.03.2024 11:10
  • 13.03.2024 11:53
  • 13.03.2024 10:22

Sorry for not specifying the seconds, we do not have that saved.

1 Like

Hi, @kristian!
Did you find any issues on your part or is it something we need to look into?

Hi!

Sorry for not responding. We have found some improvements regarding the scalability of the PDF generation service. Those are in implementation, so this will hopefully be improved soon!