How should impossible analyses be presented?

Right now, the tool only shows analyses that can be performed with data from currently open traces. Consequences:
* Users must know exactly what data is required for the analysis they want to perform.
* Users could assume a missing analysis is not implemented.

It seems valuable to be able to search through all implemented analyses

Reasons an analysis is not possible:
## Analysis missing required data
Ex. Missing event type, Missing userspace trace

* **Only show analysis in list if explicitly searched for?**
* Users could assume the analysis is not implemented
* Caters to experts (beginners may not know what to search)
* More typing/navigation work?
* **Show in analysis list, but faded?**
* Would there be too many irrelevant list items? Is there a way to guess which analyses could be relevant even if currently data is missing? Ex. Show only CPU relevant analyses if there's a kernel trace, don't show GPU analyses.
* **Present full analysis catalogue somewhere else?**

Also useful:
* **Say what data is missing for the analysis to be performed?**

## Analysis not implemented by server
* Analysis shouldn't appear in list?

The issue names no files, tests, or entry points. First review the existing analysis-list and capability handling to establish the current behavior, then seek a maintainer decision on the presentation policy; done requires an agreed UX for unavailable and unimplemented analyses, including missing-data guidance.

Written by the indexing model from the issue text.