This bug was today raised by our users. I am hoping to see a fix vey soon.

We're getting somewhat the opposite - Live Search and advanced search are not returning our UNarchived pages found. Until we include Archived in the advanced search. 

I've started coming across this in my team's Confluence instance. We're replacing a whole section of documentation with new pages in a new structure, but the old archived pages keep showing up in search results and causing confusion.

This is really unhelpful - I agree with the OP that archived pages should be excluded by default from search results, regardless of recency of view.

A nice to have might be a toggle to manually include archive pages in the results on a search-by-search basis.

So, I made an error in my earlier assessment - I don't think our caching of recent items is the issue here. Instead, the API called by search to fetch items for pre-query and use in "cached search" also fetches both trashed and archived pages. The API returns 200 items, so if the page is in your last 200 read it will show up - this explains what Mitchell described and unless you read pages very fast will last longer than 15 minutes.

We've got a few options from here - extend that API to offer archived filtering options, or switch to another API.

I understand there is some technical implications. However, this lasts a lot longer than 15 minutes on the user side. This is a constant for us, any time a query is ran through quick search it returns all pages: archived and non-archived. 

This results in the quick search being avoided by users. It also renders the archive function to be a lot less useful, if not at all.

In a large scale confluence instance the goal for a good end user experience is to be able to quickly and easily find the content that you are looking for. If it becomes too technical you immediately lose buy in for the majority of people. 

Interesting. This happens because of the caching done by the search component, which lasts for 15 minutes. I agree with the sentiment though - archiving a page is to get it out of the way, so having it show up in search is rather detrimental. Also worth noting that only cached pages can be searched in this way, the backend does not return archived pages.

I've assigned this one to Alli for now, however this will probably require a compromise between teams. Reducing the caching duration could be an option, but that will come with extra load to fetch recent items more often (longer loading times for the dialog itself). Another option could be to attach archived status to pages once they come back from the backend, but that requires and additional loading state + the potential for jank.

This bug is directly affecting our user experience. 

We have a vast Confluence database and many departments/teams working within our corporate structure. 

It makes it increasingly challenging to search for pages when archived pages show up as well. 

This confuses the user and it defeats the purpose of archiving. We archive pages because we don't want to delete them, but also we don't want people to see archived pages or at the very least include an option in the space settings to make archived pages viewable or not based on a permission.