Scenario: A larger node (in terms of data it holds) goes down. You realize that it will take slightly more than `max_hint_window_in_ms` to fix it. You have a the disk space to store some additional hints.
Proposal: Support changing `max_hint_window_in_ms` at runtime. The change doesn't have to be persisted somewhere. I'm thinking similar to changing the `compactionthroughput` etc. using `nodetool`.
Workaround: Change the value in the configuration file and do a rolling restart of all the nodes.