Call REST Endpoint execution limitation

[phertzog]

Hello,

In one of my integration I use a Call REST Endpoint to send a quite large Json payload (but under the 20 Mb limit).

Each time a call the REST endpoint, I get an error at 180s execution time. The error message is following : Request to https://xxx/mass_update_stock failed with reason: TIMEOUT_WEB

I checked the known limitation for the Call REST Endpoint task (https://cloud.google.com/application-integration/docs/quotas) and the documentation only mentions limitations on concurrent calls or message size. The document gives a 3 minutes timeout for connector calls, but this is not my case.

So my questions are : 

- Is there alse a 3 minutes timeout limitations for Call REST Endpoint task?

- Is there a way to increade this timeout?

Regards

[greb · Accepted solution]

Hi @phertzog,

I understand how annoying it can be to face timeouts, particularly when working with large payloads and critical integrations. Please see my answers to each of your key questions below:

1. Is there a 3-minute timeout limitation for the Call REST Endpoint task?
   Yes, the Call REST Endpoint task has a hardcoded 3-minute timeout limit. This is consistent with the quotas outlined in the GCP Application Integration documentation. You may identify this confirmation specifically by going through Usage limits.
2. Is there a way to increase this timeout?
   Unfortunately, no. The timeout is fixed and cannot be extended for the Call REST Endpoint task.
   

Recommendations:

1. Optimize the Backend: Check if the backend can be optimized to respond within 3 minutes.
2. Use Asynchronous Processing: Trigger the task and handle responses later with tools like Pub/Sub or Cloud Tasks.
3. Consider Alternatives: Use something like Cloud Run, which supports up to 60-minute timeouts, for long-running tasks.
4. Adjust Payload Size: If the payload is near the 20 MB limit, try splitting it into smaller chunks to reduce processing time.

I hope the above information is helpful.

[phertzog]

Hello, 

Thank you for the confirmation of the limitations. I was expecting that there is no way to bypass the 3 minutes mark, but was worth a try 🙂

I will take into account your suggestion (I already had some of these in my mind).

Regards

Regards