Baur

Hi @seut, could you please provide an early feedback? No docs/tests on this stage. As discussed, a new field in the request is needed - to differentiate whether it's manual...

> As a user you probably also don't want to impact query performance too much if you invoke it manually. if we extend grammar, user can do `throttle = on`...

Hi @seut, @mfussenegger could you please have another look? It's still a draft, no docs/tests, would start with polishing the PR once general approach is correct. TODO: restrict triggering `ANALYZE`...

> TODO: restrict triggering `ANALYZE` if scheduled one is already in progress. this is not related to throttling and will be done in a follow up PR

retest this please

> Added some doc & test suggestions I pushed a fixups https://github.com/crate/crate/pull/13046/commits/6edc3517ea035ab63697a33e68e3f1a29ca189dd, https://github.com/crate/crate/pull/13046/commits/c6980ac2304804b35eac5e46ef676fb9fb36fcde - I think we cannot refer to not yet merged doc?

> > I think we cannot refer to not yet merged doc? > > Of course you can, you can (and also please do) validate your documentation/CHANGES entry changes by...

retest this please unrelated failure https://jenkins.crate.io/job/CrateDB/job/crate_on_pr/4009/testReport/junit/io.crate.integrationtests/GroupByAggregateTest/test_select_distinct_with_limit_and_offset_applies_limit_and_offset_on_distinct_resultset/ similar to https://wacklig.pipifein.dev/github/crate/crate/c6173120-52cb-4ff6-876f-008c08ab3064/0c331b92-25ae-4493-a10c-97d453040390 looks like a new flaky but has to be addressed separately

Since weeks is mentioned as example use case: CrateDB: `SELECT 'P10D'::interval` returns `1 weeks 3 days 00:00:00` PG13: `SELECT 'P10D'::interval `returns `{"days":10}` it's just representation in client but currently internal...

Hi @mfussenegger, could you take a look? On the last fixup: Before https://github.com/crate/crate/pull/12754 it was not necessary to merge with RECOVERY before merge with UPDATE but now it's needed to...