Piotr Sarna

Since it's also part of https://github.com/scylladb/scylladb/pull/11351, I think we should close this issue and continue with https://github.com/scylladb/scylladb/pull/11351. The period in which wasm-based UDFs are disabled should be minimized/removed, and the...

Note: I only tested it semi-manually (aside from the regression tests, that all existing alternator tests continue to pass), by running a modified alternator pytest and then changing the default...

I marked the PR as "Refs", because it only "fixes" our previous incorrect assumption, that setting a default_time_to_live parameter in an alternator table via CQL is a viable workaround for...

Now that I think of it, perhaps an alternator tag (like, `system.scylla.alternator.default_time_to_live` or some other equally unique-enough name) can serve as such configuration. It's already per-table and we can simply...

our existing reserved name for the write isolation tag is `system:write_isolation` - with this scheme, `system:default_time_to_live` sounds like a right thing to use

I just converted the series to draft, because I added experimental, untested changes which introduce a new tag - `system:default_time_to_live`, which changes the value of `default_time_to_live` parameter in the internal...

> Looks good, although, as @eliransin noted, it's not nice that we don't have any tests - not even for error cases. For the success case, we can write a...

v2: - deduplicated the code by moving common procedures to helper functions - dropped the redundant and incorrect double schema updates in favor of using a single announcement, which updates...

v2: - added more paragraphs to docs, further explaining the differences between tag-based TTL and DynamoDB's notion of TTL - simplified the code for removing default TTL - changed the...

It looks like this pull request contains tons of unrelated commits due to next being repushed. Could you rebase on top of master and make it the target branch? That...