Hi, so I’m scheduling tasks to run on machines potentially hours from that moment (I’m using device farm, where I schedule large numbers jobs to run on the phones as fast as they can, but which may take a long time to get through all of them). These tasks still need certain privileges, e.g. uploading to S3 buckets, using SNS, and more. However, at that point hours later, pre-signed URLs that were generated during task creation may be expired and role chaining will definitely be expired. Short of creating new endpoints, is there anything else I can do to run these operations hours after the task is created?
Why not stop using pre-signed urls?
If you know it’s a matter of hours, the cap for timeouts on signed S3 URLs I believe is 7 days, ideally you can change your implementation of the signer so that the expiry is in (for example) 24 hours?
The cap is lower in this case (8 hours i think?), depends on the context in which the presigned url is created, and this is a lambda creating it
some mechanism is needed to upload the results of the task to S3 or elsewhere, so pre-signed urls are one option