It would be nice if eventsd could check whether the worker has completed its job and otherwise (temporarily) store the event for later processing.
Workers can fail (at least) due to:
A bug in the code resulting in an exception
An issue with an upstream system (e.g. (local) MySQL/Redis or remote services (REST API)
It would be nice if eventsd could store the event in such cases and reprocess them at a later date (with a configurable expiry and a filter which tags are important to be kept/retried (unless this applies to all)).
This would probably also require a change in workers themselves which should return the status of the _store and send events.
A (temporary) failure in the worker should not longer result in lost events.
It would be nice if eventsd could check whether the worker has completed its job and otherwise (temporarily) store the event for later processing.
Workers can fail (at least) due to:
It would be nice if eventsd could store the event in such cases and reprocess them at a later date (with a configurable expiry and a filter which tags are important to be kept/retried (unless this applies to all)).
This would probably also require a change in workers themselves which should return the status of the
_storeandsendevents.A (temporary) failure in the worker should not longer result in lost events.