I'd like to replicate an index to a remote cluster, but with the source documents modified by either masking or removing a small selection of fields.
Combining CCR with an pre-processing pipeline / ingest node looks like it would be ideal, but as far as I can see this isn't supported.
An alternative approach would be to combine an incremental reindex from remote with an ingest node, calling
_reindex on a regular basis e.g. every 5 minutes and specifying a query that would return only the documents ingested since the last call to
Does this approach seem reasonable? If so, are there any edge cases I should consider, or are there different approaches I should consider?