James Arthur
James Arthur
Yup, that’s good. There’s also the control flow issue that atm, the ShapeStream just stops on a 4xx status. So I believe a Shape or a useShape with a `headers`...
If we want to "restart" by creating a new stream, then this will help: https://github.com/electric-sql/electric/pull/1989
I think this can work nicely. Avoids having to manually restart or do the “auto-retry calling headers function again” approach — this is more explicit, less magic. Being in the...
Just an observation that a client may be able to re-connect after receiving a response within the buffer window. So we should serve new requests from memory where possible as...
Note, please wait for @KyleAMathews to comment and approve/revise/reject the approach before implementing.
Hmm, yes. So there is a downside of long duration caching of live requests: it results in iterating through lots of requests from the file cache often without data. This...
Hey @luke-stdev001 that would be very welcome. We do need to port over / add back in deployment instructions to the new site. If you were interested in taking that...
Oooh, that’s great!
In the Hammer time doc https://electric-sql.slab.com/posts/hammer-time-h8utin6y we have goals to test and demonstrate very low ongoing resource usage and how this is affected by: - number of shapes - write...
Is this only a problem if you have a compaction boundary mid transaction? If compaction always compacts to the end of a transaction and anyone requesting a previous offset gets...