See PHI251. This requests an "Ignore generated changes." option for Owners. There may not be a direct path to this today, but this is almost certainly a plannable feature.
See PHI285. An install would like a way to monitor changes to Herald rules. This mostly needs planning. The actual implementation may not be difficult, it's just not clear if "Herald rules for Herald rules" are the best approach.
See PHI592. A diff with 600MB of videos required more than two hours to turn into an email patch, which exceeded the task lease duration and cascaded into various other kinds of problems. This is somewhat adjacent to T11767.
See PHI580. An install would like write API access to object relationships (depends on, fixes, etc). This is complicated, but perhaps plannable, and we may be able to put a reasonable facade with partial support into the upstream.
See PHI483. The available actions from "Changes Planned + Draft" (and maybe "Abandoned + Draft") are currently too permissive. Fewer actions should be available; particularly, "Accept" should not be available and the actions should generally be more similar to the actions available from "Draft".
Internally, it would be nice to better document the Phabricator query layer.
See PHI604. See also T13126. An install is hitting the "large blame does not degrade gracefully" issue mentioned in that task, and would also like a way to selectively disable blame (which I was leaning toward implementing but don't think got written down anywhere).
See PHI598. See T13110. An install hunted down some specific performance issues with very large changesets in Differential. At least some of the big ones are likely trivial to get rid of.
See PHI590. See T13128. The policy tag in the header of Phriction documents could be more useful than it currently is. This is potentially the tip of a nest of vipers, but improving the Phriction behavior narrowly would be a good first step.
See PHI609. An install would like a way to filter commits by a particular branch, plus normal filter criteria. This is somewhat difficult in the general case (see the Maniphest fulltext search join issue from prior to Ferret: it's hard because we can't apply the branch constraint in MySQL, so we have to intersect "stuff on the branch" with "stuff matching the other conditions" in the application) but desirable, and likely tractable in reasonable cases.