* Fixed order
* Added missing changlog entry for GH-10145
* Fixed misclassified entry for GH-10158

Automatically populate `CONSUL_HTTP_ADDR` for connect native tasks in host networking mode.

Reserved Cores [1/4]: Structs and scheduler implementation

* website: use global featured-slider component

* website: delete unused local featured-slider

Automatically populate `CONSUL_HTTP_ADDR` for connect native tasks in host networking mode. Fixes #10239

autoscaler plugin docs

The OTT feature relies on having a query parameter for a one-time token which
gets handled by the UI. We need to make sure that query param is preserved
when redirecting from the root URL to the `/ui/` URI.

Co-authored-by: Kris Hicks <khicks@hashicorp.com>

This closes #10146.

Because of cibernox/ember-power-select#1203, which documents
the current impossibility of attaching test selectors to a
PowerSelect invocation, this uses test selectors on parent
containers instead, occasionally adding wrappers when needed.
I chose to leave the existing test selectors in the hopes that
we can return to using them eventually, but I could easily
remove them if it seems like extra noise now.

Presumably for the same reason, @class no longer works, so
this adjusts the scoping of global search CSS to preserve the style
of the search control.

I also included an update to the latest version of
ember-test-selectors, since we were far behind and I tried
that before finding the aforelinked issue.

Finally, this replaces ember-cli-uglify with ember-cli-terser to address
production build failures as described at ember-cli/ember-cli#9290.

periodic: always reset periodic children status

api: add Allocation client and server terminal status funcs.

* allow setting an enterprise license environment variable

* update comment

* address pr comments

UI: Showed host reservations and allocation utilization by task on existing stats charts

Fixes a bug where Nomad reports negative or incorrect running children
counts for periodic jobs.

The periodic dispatcher derives a child job without reseting the status.
If the periodic job has a `running` status, the derived job will start
as `running` status and transition to `pending`.  Since this is
unexpected transition, the counting in StateStore.setJobSummary gets out of sync and
result in negative/incorrect values.

Note that this only affects periodic jobs after a leader transition.
During the first job registration, the job is added with `pending` or
`""` status. However, after a leader transition, the new leader
repopulates the dispatcher heap with `"running"` status and triggers the
bug.