use scalar prob for quantile/tail ess

[OriolAbril]

I added a call in ess to the prob validating function in arviz-stats in a recent PR.
For local ess it does a loop over the given probabilities and validates each but for
tail/quantile it is called directly. The validator doesn't allow single element list/array-like
and there were some cases in plot_convergence_dist` that could use that, this should make
sure a scalar is used. Being a single case where this happened I thought it was better to fix
here than to extend the validator in arviz-stats.

Edit: forgot to commit the sequence side of prob validation, arviz-devs/arviz-stats#109 combined with this should fix it

---

📚 Documentation preview 📚: https://arviz-plots--240.org.readthedocs.build/en/240/

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 68.38%. Comparing base (352af1a) to head (15d1868).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #240   +/-   ##
=======================================
  Coverage   68.38%   68.38%           
=======================================
  Files          40       40           
  Lines        4551     4552    +1     
=======================================
+ Hits         3112     3113    +1     
  Misses       1439     1439           

☔ View full report in Codecov by Sentry.
📢 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.