Sure, there are downsides.
1. The replay of WAL records becomes somewhat more CPU-expensive. But at the same time, OrioleDB can do that in parallel.
2. Replace becomes not the file-level equivalent of primary. Thus, you, for instance, can't continue file-level incremental backup started taken from the primary with the replica.
Nevertheless, I strongly believe that row-level WAL is a great win on average, and opens brilliant perspectives for further developments.
Comments
What is the downside of doing this? I assume there is a tradeoff?
Sure, there are downsides. 1. The replay of WAL records becomes somewhat more CPU-expensive. But at the same time, OrioleDB can do that in parallel. 2. Replace becomes not the file-level equivalent of primary. Thus, you, for instance, can't continue file-level incremental backup started taken from the primary with the replica. Nevertheless, I strongly believe that row-level WAL is a great win on average, and opens brilliant perspectives for further developments.