I suspect that after T387210: Latest log query timeout due to ill-advised indexing the log queries are going to be notably faster, and will perhaps be doable. Looking at what could go here:

• general statistics
  • out of "basic info" registration date, first/latest edit, latest log, num edits (including deleted). perhaps thanks received. not user id as they're local
  • then there's the question of how we handle the "last 5000". 5k per wiki? could be too much. the last 5k overall? except by selecting the whole tables and unioning them (which defeats the purpose), I think we can't query from the rows of this_p.revision and the rows of that_p.revision. Or, we could do the last 10 or 20 or 100 or whatever (TBD) per wiki. shouldn't be too slow
    • aside from that all of "edits (live)" can be reused AFAIK. Well, except the links to other project-specific tools
  • All the rest of should be reusable.
  • we would be adding a global actions column, too.
• the charts: theoretically we could keep all; discarding namespace and instead keeping wiki. Looking at the queries all should be adaptable; essentially just ignoring namespaces and repeating for each wiki.
• topedits: probably shouldn't be included. We've already got a long lot of tables (btw, why not flex these? saves vertical screen space); multiplying by wiki would make it prohibitive. Plus, the query time.
• local rights changes wouldn't make much sense, but I'd argue global right changes should be included

I'm wondering about the chart colors. If we take the values of the colorlist and reindex them with incrementing indexes (as opposed to namespace IDs), we get 85 of them. If someone hypothetically edits more than 85 wikis, which is plausible, the legend is going to make a lot less sense.

Or, if user edited >= 86 wikis overall, we could fall back to merging all wiki types together, i.e. Wikipedias, Wikisources, and so on. Only if isWMF(), I suppose.

Looking at the MW request a table of user_id, registration date, and user groups shouldn't be too costy either I suppose.

On local rights: probably a simple listing of "on which wiki is this user an admin" could be useful and not too costy.