List View Threshold (5000 items): What limitations for 5k-issue can we expect even after the "correct" columns are indexed (manually or automatically)

Question

The infamous 5000 Threshold view limit [1, 2, 3] is a well-known struggle.

One often suggested solution is to "plan ahead", and let's imagine we have done this and indexed our most important columns, or they have been autoindexed, and our views return less than 5000 items ‒ what kind of limitations will still remain?

There is an overview on support.office.com of blocked operations:

• Create or remove an index.
• Use the Open with Explorer command.
• Filter and sort based on non-indexed columns.
• Set permissions.
• Show totals.
• Save a list as a template with data.
• Other operations
• Delete a folder or list.
• Change a column or column type (except a name change).
• Enable or disable list attachments.
• Create a lookup column that enforces a relationship.
• Copy or rename a folder.

I'm not sure if all the points above applies after indexing.
I don't know if they apply to both libraries and plain lists.
I also think there might be other limitations, especially in relation to JSOM, REST or Workflows, for example:

• Filtering on multi lookup columns.
• Not all field types can be indexed.

Update:

• Filtering (with CAML) by text on lookup fields, even single lookup.

Update 2018-01-24:

• Using OrderBy on text fields in CAML, even when fields are indexed (CSOM, JSOM, REST). $orderBy in OData works.

Update 2019-01-03:

• Setting multi lookup values when target list is a library.

Answer 1

Once over 5000, REST becomes the best method I've found so far to programmatically manipulate data. It will take your request and return data in batches of 200, if more items satisfy your condition, it will give you a new URL to call to get the next page of data. I've used this to batch update a library of 150K+ files with no trouble.

Everything else there is correct from my experience. Some other things that might possibly be limited in functionality are Information management policy settings and Information Rights Management.

Honestly though, I can't reiterate how important it is to properly plan for situations like this. You need to know what you are storing, how to classify it, and derive a folder structure that will assist in keeping you compliant with view limits. The folders can be configured to automatically set metadata (column default values which will help with columns having data for indexing purposes).

Understand file retention limits and create Information management policies to expire documents when they reach those time frames. Create a custom search page scoped to your document library with preconfigured refiners based on your data.