Commands that specify a string value of 

"_id_"

 (the name of the id index) for the "hint" option succeed when they should fail on timeseries collections. Timeseries is a type of clustered collection. Even though clustered collections do have an index on _id by default, for timeseries, that _id index is on the buckets coll and not the user-facing coll so these types of commands with hints on _id should fail for timeseries.

Ex: 

db.adithiTS.find({x: {$gt: 2}}).hint("_id_")

db.adithiTS.deleteMany({x: {$gt: 2}}, {hint: "_id_"})

or 

db.runCommand({ find: "adithiTS", hint: "_id_" }) 

db.runCommand({ delete: "adithiTS", deletes: [{q: {x: {$gt: 2}}, limit: 0, hint: "_id_" }]})

I only tested the find/delete commands but this likely affects all commands that have a "hint" option (ex: bulkWrite, findAndModify, update, etc.). The inclusion of the hint: "id" option doesn't appear to affect the expected results in my preliminary testing.

These commands (with string hint: "id") pass on the latest stable versions for 6.0, 7.0, 8.0, and on master. When the hint has an object value, like hint: {_id: 1}, the commands fail as expected with 

planner returned error :: caused by :: hint provided does not correspond to an existing index