fix(prometheus)!: default group_path to True to bound metric cardinality

[Bishwas-py]

Description

Fixes #4891.

PrometheusMiddleware records the path metric label from the raw request
path by default. On a route with path parameters, every distinct value seen
in traffic becomes a new time series, so the metric set grows without bound
as traffic hits distinct parameter values, for example /users/1,
/users/2, /users/3, ... each get their own series.

group_path=True already avoids this by using the route template
(/users/{user_id}) instead of the raw path. The mechanics of that option
were made robust in #3533, #3687 and #3784. The remaining gap was that the
safe behavior was opt-in, so users hit the cardinality trap silently by
using the default.

Fix

Flip the default to True. Raw-path behavior stays fully available via
group_path=False for anyone who wants it. Also updated the docs example's
local default to match, so the snippet shown in the Prometheus docs page
reflects the new default rather than silently diverging from it.

How was this tested

Added two tests in tests/unit/test_plugins/test_prometheus.py:

• PrometheusConfig().group_path is True
• a route with a path parameter hit with three different values produces a
  single path="/users/{user_id}" series, not three separate raw-path
  series

Both fail on main (raw paths leak through) and pass with this change.

Also ran the reproduction script from the issue directly:

group_path=False: 5 distinct app path series -> ['/users/1', '/users/2', '/users/3', '/users/4', '/users/5']
group_path=True:  1 distinct app path series -> ['/users/{user_id}']

With this change, calling PrometheusConfig() with no arguments now
produces the group_path=True result above instead of the False one.

Full suite: tests/unit/test_plugins/ and tests/examples/test_contrib/prometheus/ pass (491 tests), no existing test asserted raw-path grouping since none of them use parameterized routes.

---

📚 Documentation preview 📚: https://litestar-org.github.io/litestar-docs-preview/4925

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 67.30%. Comparing base (2453f45) to head (cf7c82a).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #4925   +/-   ##
=======================================
  Coverage   67.30%   67.30%           
=======================================
  Files         293      293           
  Lines       15230    15230           
  Branches     1728     1728           
=======================================
  Hits        10251    10251           
  Misses       4832     4832           
  Partials      147      147           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.