This mitigates the impact of an ambiguous case we can't quickly fix
properly; see #27521 for the crash and #36185 for the larger issue.

With this change the impact of #27521 is reduced so it is not blocking
48.

* Run ReplayIO E2E tests every three hours

* Fix the conditional and run all tests

* Upgrade replay runner to the latest version

* Use unique GitHub run id to tag Replay runs

* Make replay API key available to each run

* Use native replay upload

* Upgrade replay plugin to v1.7.5

* Attach dashcard's viz settings to the card's :data :viz-settings key

This flows through to the export format fn `metabase.email.messages/stream-api-results-to-export-format`
Where it is then properly used.

* Added test that gets the contents of .csv attachments

Not sure yet why it's not working. Locally, with a real db id it works as expected, so it's on the right track.

* Create a working test.

* Card qp merges dashcard viz when it exists. Viz middleware normalizes card/dashcard viz-settings

* Addressed feedback.

 - if viz settings map is empty in process viz settings middleware, use qp store.
 - now we don't have to do the (or ...) logic anywhere else
 - added comment in the test to try clarify what the 'viz' is setting up

* Mark pivot tables test that failed recently

A failure in `master` from November 23, 2023.
https://www.deploysentinel.com/ci/runs/655f74c4fcacf3cead2261a5

* Tag top 10-ish recent flakes

* Remove default exports from dashboard/*
* Remove dashboard index.js/ts re-exports 

* Update athena.md

Added notes about permissions requirements especially re: Lake Formation as a source and changed headings to have a H2 permissions heading with the example bucket policy under it.

* Update docs/databases/connections/athena.md

Co-authored-by: Jeff Bruemmer <jeff.bruemmer@gmail.com>

---------

Co-authored-by: Jeff Bruemmer <jeff.bruemmer@gmail.com>

support Percentage setting in the semantic type editor, infer percentage column formatting default (#36186)

* Adding Percent Metadata Type

This adds the :type/Percentage metadata type to Metabase. It does
not currently do any fingerprinting. It just enables setting of the
type for the FE.

* Making :type/Percentage a semantic type as well as a decimal.

* support Percentage setting in the semantic type editor, infer percentage column formatting default

* Support for static-viz rendering when user metadata is `:type/Percentage`

This PR adds in user-defined metadata in `prep-for-html-rendering` so that `number-formatter` can properly render the column as a percentage. This may actually fix a family of issues as user-defined metadata doesn't appear to be used anywhere in this ns. We still need to consider cases where the formatting is specified by the viz settings, but this one step in the right direction.

* Percent semantic types render as percents in tables

This PR adds logic to properly capture a percent semantic type and renders it as such in a table static viz. Note that the `number-formatter` function in `metabase.pulse.render.common` has logic that renders the percent as a truncated int. IDK if this the desired long term behavior. I'll have to ask product that, but if we just make the change it could have potential impacts on places that expect this behavior. It seems like it would make more sense to be like Excel and allow each column to set its significant digits.

* Revert "support Percentage setting in the semantic type editor, infer percentage column formatting default"

This reverts commit e001e6312e6f0bd5700c1d9eb28c88829d87a70a.

---------

Co-authored-by: Aleksandr Lesnenko <alxnddr@gmail.com>

The fix in #35895 included reducing the size of the printed object, but the drilled elements needed to be pulled from the object key in the event, not the event itself. This fixes that.

* `can-write?` dashboards + cards in Custom Reports

* use the audit-db-id which will always be there

* in tests, audit collection should have id 13371337

- and type = 'instance-analytics'

* throw in type just for fun

* Update test/metabase/models/collection_test.clj

don't use the magic number

* better default collection detection

---------

Co-authored-by: Ryan Laurie <iethree@gmail.com>

* Update models.md

added a note about the timezone used for model caching cron schedules

* Update docs/data-modeling/models.md

Co-authored-by: Jeff Bruemmer <jeff.bruemmer@gmail.com>

---------

Co-authored-by: Jeff Bruemmer <jeff.bruemmer@gmail.com>

This replaces the existing setting max-results-bare-rows with two settings:

- `max-unaggregated-query-row-limit`, and
- `max-aggregated-query-row-limit`

Co-authored-by: Noah Moss <noahbmoss@gmail.com>
Co-authored-by: Ryan Laurie <iethree@gmail.com>
Co-authored-by: Noah Moss <32746338+noahmoss@users.noreply.github.com>
Co-authored-by: Luiz Arakaki <luiz.arakaki@metabase.com>

* Update upgrading-metabase.md

added note about major version upgrades taking longer

* Update docs/installation-and-operation/upgrading-metabase.md

Co-authored-by: Jeff Bruemmer <jeff.bruemmer@gmail.com>

---------

Co-authored-by: Luis Paolini <paoliniluis@gmail.com>
Co-authored-by: Jeff Bruemmer <jeff.bruemmer@gmail.com>

This reverts commit 210978db.

* update yamls for latest change on key order

* bring custom reports back

* add Metabase Analytics instance type

* add new view v_tasks, iterate content adding links

* add v_tasks to permissions

* fix double precision types for h2 and mysql

* add type to instance analytics col

* add mysql rollback

* fix paths in migration

* handle dynamic tab ids

* update yamls click behavior

* add namespace=analytics to audit collection

* add type to collection

---------

Co-authored-by: Noah Moss <noahbmoss@gmail.com>
Co-authored-by: Ryan Laurie <iethree@gmail.com>
Co-authored-by: Noah Moss <32746338+noahmoss@users.noreply.github.com>

* Unskip previously quarantined pie chart nulls test

* Update card details

* QueryBuild and SQL questions: hide Explore results with parameters
* Fix AtomicQuery.canNest return type (so that it is always boolean)
* Add Question.canExploreResults tests

Co-authored-by: Ngoc Khuat <qn.khuat@gmail.com>
Co-authored-by: Cal Herries <39073188+calherries@users.noreply.github.com>

adds new `ascendants` multimethod, which traverses tree upwards w/o looking for children.

Pulse and related are skipped for now: they have a full serialization support, but are not serialized.

* Adjust mysql container env

* Setting the HOST can cause trouble they said

https://hub.docker.com/_/mysql. Let's try it without then.

* Also MYSQL_USER doesn't play well with empty password

* Cast results of temporal extract to decimal

* Update test

* Update test

* Revert casting temporal-extract results to decimals

* Handle mysql YEAR type as a number instead of a date

* Cleanup

* Use mongo sessions in aggregation pipelines

Using sessions, cancel signal from streaming response can be handled
correctly - query is killed also on database side.

* Update tests

* Let the middleware handle the exception

While coding the other test, I've found out that if exception is handled
in the `reducible-results`, code clogs for 3 minutes,
which is `.maxTime` for aggregation. I don't know the exact reason yet,
but for time being I'm leaving the exception handling to middleware.

* Add testing string

* Encode aggregation results as `BasicDBObject`

Use this type instead of BSON's Document because code transforming
DBObject to clojure structures is already in place.

* Update exceptions handling and tests

Exception handling now more resembles code in other drivers. Tests were updated
to use mbql query instead of raw native to avoid differences between various
mongodb versions. `$dateTrunc` was not available in version 4.

* Increase cancelation interval in tests

* Address remarks and update test to handle cold start

By cold start I mean running it before dataset is initialized.

* Remove unnecessary hint