Support hourly index patterns

[jmacdone]

Description

Collect otherwise missing data for top_count_keys for those indexing hourly with a %H pattern, eg. logstash-%Y.%m.%d.%H Previously format_index() assumed daily was the most granular indexing.

See also https://github.com/jertel/elastalert2/discussions/1326

Checklist

• [x] I have reviewed the contributing guidelines.
• [x] I have included unit tests for my changes or additions.
• [x] I have successfully run make test-docker with my changes.
• [x] I have manually tested all relevant modes of the change in this PR.
• [ ] I have updated the documentation.
• [x] I have updated the changelog.

Questions or Comments

test-docker

py311: OK (575.05=setup[483.55]+cmd[89.54,1.97] seconds) docs: OK (481.68=setup[466.33]+cmd[15.35] seconds) congratulations :) (1056.81 seconds)

[jmacdone]

• committed @jertel 's suggested changes with fix-ups
• updated the unit test to reflect intended meaning of the add_extra param
• rebased from master