Hi Everyone ,
I hope this message finds you all well.
I’m trying to understand the current state and intention of the ‘changefeed’ functionality. I saw an update last year on this and saw it was noted there are limitations.
Using this inside the fdbcli it seems to match what I’m looking for, it would be great if someone could help me understand:
-
Implementation Limitations: What are the high level limitations? e.g. Limits on retention period, a limit on the number of supported consumers, issues with scaling across multiple consumer groups, or any other constraints? Are there performance or scalability concerns? etc
-
API Binding: Is there an intention to surface the feature in the API bindings? After exploring the source I couldn’t see any implementation outside of the fdbcli directly.
-
Use case fit: Assuming binding were in place. Would this be the best way to keep an external analytics system in sync with fdb? Or is it more appropriate to model a changefeed in the tuple layer using watches etc?
Thanks in advance!
James