Recently accessed items do a full page reload

[flash1293]

Kibana version: master

Describe the bug: When clicking the links in the "recently accessed" list of the side nav, a full page reload is performed. The navigation to these items should use the flow of application.navigateToApp if applicable to perform the navigation to avoid reloading the page (basically in the same way the actual nav links work).

This is currently not a problem because most features using this feature are in the legacy platform in a single "app" anyway, but when cutting over to the new platform, this won't be true anymore.

Not sure whether this is a Core UI or a platform concern.

[elasticmachine]

Pinging @elastic/kibana-core-ui (Team:Core UI)

[Dosant]

I think, ideally, this needs: #58751

[myasonik]

Recently accessed items will also be going away soon so not sure if this is worth the effort... @ryankeairns thoughts?

[flash1293]

If it’s on the way out I’m +1 for not putting any effort in it - if #58751 fixes it, good, otherwise it’s not a big deal. The list is still functional, just not the perfect experience

[ryankeairns]

The plan, as of right now, is to display recents in the Global Search feature. We will keep an eye on this one and see if the experience becomes intolerable in the meantime.

As part of the Kibana home page/landing page project there is a chance we also provide an in-page solution again. It's all on my radar, thanks for pointing out the degraded experience!

[elasticmachine]

Pinging @elastic/kibana-platform (Team:Platform)

[pgayvallet]

@ryankeairns

The plan, as of right now, is to display recents in the Global Search feature. We will keep an eye on this one and see if the experience becomes intolerable in the meantime.

Just to be sure as it's quite old now: #72331 still have the recently accessed items in the left nav, right? Just to confirm this issue is not outdated.

[ryankeairns]

@pgayvallet

The current thinking is that we ship search as-is, leave recents in the left nav, and collect customer feedback before determining the 'ideal' default state.
A separate issue was created to track the factors and suggestions which led us to this approach: #75040