* wip

* fix caching behavior for audit db in database-id->connection-pool

* test that caching behavior works as advertised

- db->pooled-connection-spec always returns the same pool for audit-db
- audit-db-id is not in the database-id->connection-pool cache

* DISABLE audit db exposure through GET api/database

* Lots of things

- automagically loads instance analytics on app startup
- TODO: logging in after that is bokred

* add some logs / fix comment

* remove internal_analytics resource

* fix linter issues + update logging

* add tests for existing and missing content

- add test replica of content into test-resources

* remove unused require

* delete sample db + personal collection from test-resources

* get instance_analytics resource off the resource path

* point it to the proper IA resource

* un-hotwire the change to show audit DB on GET /databases

* fix test + lookup the right resource

* remove creator_id references from instance_analytics export test data

- get test feedback

* Revert "remove creator_id references from instance_analytics export test data"

This reverts commit 1aa9d4b1dad304ac0c6d88221f81432d78fc88f0.

* fix tests

* lint

* fix tests

* more test fixing

* linter fix

* revert yarn.lock

* Look up the correct instance_analytics resource

* cleanup comment in serialization/cmd.clj

* silence noisy serialization import logs

* remove a bunch of unused serialization files

* rename Audit Database + docstring

* rename audit database everywhere

* Put the database and tables back

* constrain tests to postgres

* ignore ia data in a test

* Question URL methods to functions

* Move URLs to their own file

---------

Co-authored-by: Anton Kulyk <kuliks.anton@gmail.com>

Updates Snowflake driver to decode uploaded private key file data if necessary

* initial commit

* add require

* update gen hash

* fix test

* add email logic to alerts

* fix one test

* test fix

* re-add retry

* fix test

* add test

* address comments

* fix typo

* fix test

* fix test

* update req

We stopped using Currents.dev analytics a few months ago.
This PR removes the stray project ID from E2E workflow.

[ci skip]

* add `TabEmptyState` for empty dashboard tab

* convert files to typescript

* fix visual issues on dashboard tabs

* remove unnecessary focus css selector

* remove unnecessary position relative

* go back to using flex instead of inline-block

* Fix `Card.tsx` padding

This commit fixes a visual regression in the `Card.tsx` component,
which happened after we removed the styled system in #31306.

* fix automatic dashboard sidebar cards

---------

Co-authored-by: Kyle Doherty <5248953+kdoh@users.noreply.github.com>

* hydrate orders-tabs for public/embed dashboard

* adds test

* return dashboard_tab_id when hydrate ordered_cards for public dashboard

* show tabs in `PublicDashboard` on FE

* remove unused export

* update e2e test

* remove duplicated implementation for macro in test

* makes the macro public for testing

* run prettier on DashboardTabs.tsx

---------

Co-authored-by: Emmad Usmani <emmadusmani@berkeley.edu>

* wip

* fix caching behavior for audit db in database-id->connection-pool

* test that caching behavior works as advertised

- db->pooled-connection-spec always returns the same pool for audit-db
- audit-db-id is not in the database-id->connection-pool cache

* DISABLE audit db exposure through GET api/database

* Lots of things

- automagically loads instance analytics on app startup
- TODO: logging in after that is bokred

* Revert "Lots of things"

This reverts commit 7eb5d57b221e4a91298e9d90fc9ed3e13922aeef.

* Ignore audit-db for the "We're taking a look at your database!" modal

See discussion at https://metaboat.slack.com/archives/C5XHN8GLW/p1685964332028149

* Sync but don't scan tables from uploaded CSVs

Delete the CSV file afterwards

* Delete the CSV file more intelligently

* Scan asynchronously

never mind about the file deletion

* Turn file deletion back on

* Humanize model name for uploaded CSVs

* Update Cypress test with humanized name

* Only scan new tables, not the whole DB

* Only sync tables

* Get schema name properly

* Fix error-handling; don't create errant table

* Check for syncable schemas

* Fix E2E test

* Tidy up include-schema? usage

* Activity to toucan2

* LoginHistory to toucan2

* ViewLog to toucan2

* TaskHistory to toucan2

* PersistedInfo to toucan2

* initial changes

* fix whitespace

* update ns

* update test

* update test

* update test

* update test

Fixes #30951.

Co-authored-by: Cam Saul <1455846+camsaul@users.noreply.github.com>

* rip out styled-system usage

* fix yarn type-check errors, replace space properties w/ className equivelants

* find / replace space prop usage [ci skip]

* find / replace use of hover styled-system prop [ci skip]

* account for fontSize usage w/ Text component [ci skip]

* use font-size and weight props for styling text

* fix a few additional uses of color

* fix lint

* fix lint

* fix dumb rebase mistake

* Fix prop-types

* Add aria-labels

* Use button tag for clickable element

* Fix aria-label l10n

* Set `onClick` on a clickable element

* Add test coverage

* Remove todo

* Make NativeQueryEditor testable
- Mock just NativeQueryEditor.prototype.loadAceEditor instead of entire NativeQueryEditor

* Fix invalid propTypes

* Fix DatasetQueryEditor not re-rendering when props change
- remove custom arePropsEqual argument from React.memo() on NativeQueryEditor
- Fixes #30680

* Add a dirty initial NativeQueryEditor test setup

* Add DatasetQueryEditor tests

* Revert "Add a dirty initial NativeQueryEditor test setup"

This reverts commit 0f985db272a2e8f1be755bda07c3b642fdf90f83.

* Get rid of test errors

* Clean up DatasetQueryEditor test

* Add a test case for re-rendering of DatasetQueryEditor

* Make setup function more reusable

* Add 2 more test cases

* Bring back global NativeQueryEditor mock as many tests depend on it

* Add an e2e test for #30680

* Optimize DatasetEditor by conditionally rendering DatasetQueryEditor only when it can be seen

* Remove redundant code

* ESLint

* Use createMockCollection instead of a cast

* Move all Native Query creation to createMockNativeDatasetQuery

* Use setupCollectionsEndpoints

* Introduce setupNativeQuerySnippetEndpoints and use it in DatasetQueryEditor.unit.spec.tsx

* Extract explanatory importDatasetQueryEditor

* Remove file accidentally added when merging master

* Add missing data-testid that got lost during a merge

We migrated from CircleCI a long time ago and are now using GitHub Actions.
But even without that CircleCI, like many other CI providers, sets the
`CI` environment variable to `true` or to `1` by default.

This PR updates the `interactive?` check to use that env var as fallback.

* [MLv2] Add expressionable-columns

* Fix test