As Cloud Feeds engineers, we would like Repose's client-auth-n atom-feed feature, as described here:
to not read the entire feed in every check-interval.
Based on the code that's here:
it looks like every check-interval, the filter will get the most recent page from Identity feed, follow the previous links until there's no more previous link (i.e: you reach the end of the feed and found the oldest events that exist in Cloud Feeds). This is 3 days worth of data that it is reading over and over again.
While this may not break the intended functionality of the client-auth-n's atom-feed, it introduces unnecessary load on Cloud Feeds systems. Imagine this feature being used by all products that are currently using Repose and client-auth-n, all of those Repose nodes making HTTP GET calls this way, it would be a lot calls.
Please consult the recommended best practices for Cloud Feeds consumers here:
Please also don't hesitate to contact one of us (or firstname.lastname@example.org) if you have any questions regarding this.