* Move dashboard bookmarks spec to `organization`

* Move dashboard filter related specs to appropriate group

* Create new `dashboard-cards` E2E group

* Introduce backport reminder CI workflow

* Only trigger against `master`

* Re-trigger when author adds a label

* Disable PR bot reviews

* Keep the job name terse

based on https://github.com/orgs/community/discussions/26303 and my experience, workflows with `always()` will not cancel even if a new run of the same workflow appears with concurrency and cancel-in-progress:true. Official documentation suggests using `!cancelled()` instead

* Install Chrome immediately after checkout

* Switch to Chromium 111.0.5563.146

As of two days ago we started seeing Chrome crashing in E2E tests.
An example of the failed run: https://github.com/metabase/metabase/actions/runs/5665970958/job/15354475344

```
We detected that the Chromium Renderer process just crashed.

This is the equivalent to seeing the 'sad face' when Chrome dies.

This can happen for a number of different reasons:

- You wrote an endless loop and you must fix your own code
- You are running Docker (there is an easy fix for this: see link below)
- You are running lots of tests on a memory intense application.
    - Try enabling experimentalMemoryManagement in your config file.
    - Try lowering numTestsKeptInMemory in your config file.
- You are running in a memory starved VM environment.
    - Try enabling experimentalMemoryManagement in your config file.
    - Try lowering numTestsKeptInMemory in your config file.
- There are problems with your GPU / GPU drivers
- There are browser bugs in Chromium
```

I'm not quite sure whether this is a bug in Chrome v115 or some interplay between Cypress and this particular Chrome version, but running all tests on older Chrome version v114 seems to be working just fine.
https://github.com/metabase/metabase/actions/runs/5677579526?pr=32688

Proposal
Let's merge this ASAP as a hotfix to unblock everyone, and then we'll follow-up when there's a clear resolution of this problem (most likely upstream).

* untangle ldap circular dependency

* only try to require ee code when ee is available

* LDAP login e2e tests

* fix eslint errors

* Update e2e/test/scenarios/admin/settings/sso/ldap.cy.spec.js

Co-authored-by: Nemanja Glumac <31325167+nemanjaglumac@users.noreply.github.com>

* update command in e2e-ldap-helpers.js

---------

Co-authored-by: Nemanja Glumac <31325167+nemanjaglumac@users.noreply.github.com>

* Start all tests without a token

* Update tests

* Fix tests in `admin` group

* Fix tests in `dashboard-filters` group

* Fix tests `visualizations` group

* Batch fixes

* Fix database banner snowplow test

* Fix full-app mobile view banner test

* Simplify and remove `isPremiumActive` check

* Fix the helper

* Fix helper to fail fast

* Fix tests

* Ensure token scope is set

* Ensure CYPRESS_ env tokens are present

* Set token only if running Metabase ee

* Improve JSDoc

* Slightly tweak log message

* Update E2E dev guide

* Update cross-version tests

* Remove premium token from Cypress backend setup

* Improve a comment

* Fix test

* Update embedding copy

---------

Co-authored-by: Alexander Polyankin <alexander.polyankin@metabase.com>

This reverts commit 9da0bd76.

The previous setting accidentally executed a scheduled job **every minute**
on Sunday, when we actually wanted it to run only once on that day.

This PR fixes it and it will now run at 10PM every Sunday.
https://crontab.guru/#0_22_*_*_0

[ci skip]

* Fix Docker `latest` tag logic in the release workflow

As noted in #32036, the current logic doesn't account for the existence
of the N+1 RC tag when the current release is N. In other words, it
didn't promote `v1.46.6` to the latest Docker tag because it included
`v1.47.0-RC1` in the list of the relevant tags it compares against.

This PR fixes that and closes #32036.

**Before:**
```
Current version for v1.46.6 is {
  prefix: 'v1',
  feature: 46,
  maintenance: 6,
  build: NaN,
  isOSS: false,
  isEE: true,
  prerelease: false
}

Enumerating all git tag...
Found total 414 tags

Filtering for EE
Found total 116 filtered tags

Sorting tags to find the highest version numbers...
Showing 20 tags with the highest versions...
     v1.47.0-RC1
     v1.47.0-RC2
---> v1.46.6
     v1.46.5
     v1.46.4
     v1.46.3
     v1.46.2
     v1.46.1
     v1.46.0
     v1.46.0-RC1
     v1.46.0-RC2
     v1.46.0-RC3
     v1.46.0-RC4
     v1.45.4
     v1.45.3.1
     v1.45.3
     v1.45.2.1
     v1.45.2
     v1.45.1
     v1.45.0

The latest container image stays as v1.47.0-RC1
There is no need to tag the v1.46.6 container image as latest.
```

**After the fix:**
```
Current version for v1.46.6 is {
  prefix: 'v1',
  feature: 46,
  maintenance: 6,
  build: NaN,
  isOSS: false,
  isEE: true,
  prerelease: false
}

Enumerating all git tags...
Found total 430 tags

Filtering for EE which excludes prerelease tags...
Found total 94 filtered tags

Sorting tags to find the highest version numbers...
Showing 20 tags with the highest versions...
---> v1.46.6
     v1.46.5
     v1.46.4
     v1.46.3
     v1.46.2
     v1.46.1
     v1.46.0
     v1.45.4
     v1.45.3.1
     v1.45.3
     v1.45.2
     v1.45.1
     v1.45.0
     v1.44.7
     v1.44.6.1
     v1.44.6
     v1.44.5
     v1.44.4
     v1.44.3
     v1.44.2

Thus, the container image for v1.46.6 must be marked as latest.
```

* Remove unnecessary whitespace

[ci skip]

Replay is adding a significant overhead and is making every single run
fail. This is messing with our statistics and analytics.

While the Replay.io team is working on some optimizations that would
remove the overhead, we can significantly dial down the frequency of
these runs. We'll switch to running them only once per week, on Sunday.

For now...

Running E2E tests using ReplayIO still adds a significant overhead, which
leads to a lot of failed runs. There's no point in re-running those tests
for at least two reasons:
1. they will fail again and we're just wasting CI infrastructure
2. our slack-failure-alert-bot will spam our Slack with the reports of
the failed runs that we don't (currently) care about

[ci skip]

I forgot to do this in #31432.
As the result, some PRs have been stuck in CI due to the required check.

This PR will run the fallback job for `actions` E2E group when changed
files don't require E2E tests to run.

[ci skip]

Previously: one “feature or project” template with the `.Epic` label automatically attached.
Now: two templates, one for epics, one for smaller things that don't require milestones (or `.Epic` label; they get `.Task` instead).

Squashed commit messages follow:

* Remove unused optional field

* Update title to current convention

* Fix typo

* Remove unused section

* Change capitalization for consistency

* Remove redundant description

* Rename the issue template to mention epic

* Shorten the description of the issue template

* Lighten up the milestones section

Unfortunately, it doesn't seem possible to prepopulate the issue with an
empty tasklist (https://github.com/orgs/community/discussions/50502).

* Add an issue template for tasks [ci skip]

We had one for epics, with milestones and `.Epic` label. Pretty often
that label was accidentally left on smaller issues that were created
using the same template. Now the smaller issues have their own template
and their own label. Yay.

https://metaboat.slack.com/archives/C04CGFHCJDD/p1686548875239219

* Remove redundant angle brackets

* Use a string to get a valid yaml file

Without the quotes, github says
`Error in user YAML: (<unknown>): did not find expected key while parsing
a block mapping at line 1 column 1`

* Remove rarely if ever used placeholders

* Call out optional-ness of some links

* Contextualize a header

Context can be some text, a slack thread, a notion page, another issue,
whatever.

* Extract `questionDetails`

* Move `public-sharing` to the `sharing` folder

* Remove redundant test

This part was covered in `public-sharing.cy.spec.js`.

* Create a separate `public-dashboard` E2E spec

- Extract relevant part from `public.cy.spec.js`
- Make all tests in `public-dashboard.cy.spec.js` run in isolation

* Make tests in `public.cy.spec.js` run in isolation

* Remove redundant wait

* Limit the query results to speed test up

* Merge public questions E2E tests

* Merge `downloads` with the `sharing` E2E group

* Remove `downloads` from the folder list

* Refactor `public-dashboard.cy.spec.js`

- Clean up linter warnings by using semantic selectors
- Speed test up by using API

Stress-testing a spec with a lot of tests can easily time out before
20 minutes mark. This commit is bumping that timeout to 60 minutes.

[ci skip]

* Move core actions E2E tests to a separate folder

* Run new `actions` E2E group in CI

Metabase actions deserve their own E2E group, as the test suite around
them is expected to grow. This commit introduces `actions` E2E group to
the CI.

We're seeing increasing number of timeouts in CI without any significant
changes in the test suite. This could be due to GitHub's runners performance.

[ci skip]

Simply invoking `.bin/build.sh` will produce the uberjar without a
version. This is why @ariya created #30915 in the first place.

But the build script already takes `:version` and `:edition` arguments.
That is exactly what this PR is doing rather than manually unzipping the
uberjar in order to tweak its `version.properties` file.

[ci skip]

@jaril informed me that Replay.io is missing metadata
from Metabase E2E tests. Apparently trying to access
system environment variable from Cypress support file
didn't work.

The solution for this is to prefix the environment variable
with `CYPRESS_` and then to read it using `Cypress.env()`

This is what Cypress documentation suggests:
https://docs.cypress.io/guides/guides/environment-variables#Option-3-CYPRESS_

[ci skip]

https://github.blog/changelog/2022-10-11-github-actions-deprecating-save-state-and-set-output-commands/