Good morning everyone, when we deal with indexes we can have two use case: synchronous and asynchronous. In the first use case by being a commit hook the "handle" is returned to the application using oak only when all the hooks are completed and therefore the index is updated.
In the latter there's no way for application to know when the index is updated and therefore risking to serve stale content. Would it make sense to have an observation event on Asynchronous index update? Something like: index A has indexed nodes B, C, D. I don't think that a notification for each node is nice, I'd rather use a "batch" notification. In this way a client application by registering on such events could flush any cache mechanism that could have been implemented. Another use case would be to being able to understand about index updates on cluster. So being able to understand when an index is updated by a sychronisation of cluster nodes. If I got the above correct, WDYT about the approach? Should I file a ticket about this improvement (if not already there)? Cheers Davide
