Mitigate frequent callout failure and false determination of failure

[Michael_Summitt]

Callouts are frequently failing, and worse, detect that they have failed when in fact they succeeded.

A more intelligenty retry strategy may be helpful.  A failed callout could be retried on a staggered schedule, for example 1 minute, 3 minutes, 6 minutes, 10 minutes, 15 minutes (1 + 2 + 3 + 4 + 5).

The ability to force a failed callout to retry again from the UI might be helpful.

The ability to stop a  failed callout from retrying (without stopping all retries for all callouts) from the UI might be helpful.

[feisley]

I agree the ability to perform an on-demand callout retry would be very helpful in situations where the provisioning system was down or due to some internal error needs the callout to be retried.

Currently our only workaround is to view the parameters and send them using an external REST tool.

[lukasz]

As part of the more comprehensive Notifications feature coming soon, we're going to continue providing the access to Notification History, documented here, and are considering adding a way to force a retry at an arbitrary time for those Notifications that have failed to be delivered (specifically callouts). Hopefully, that'll resolve most of the common retry strategies for callouts.