I know vedal already said he isn't going to implement it, but as a programmer and kind of the worst guy, my brain can't help but think of solutions.
Why not have the server send payloads like:
S2C:
{
"type": "turn_on_lavalamp",
"delay_predicate": "2 hours"
}```
and
```json
{
"type": "turn_on_lavalamp",
"delay_predicate": "6 hours"
}```
and have the lavalamp turn on based on checking its internal settings against the predicate. The above message would be sent at 2 hours behind stream start, and have them send every interval of 30 minutes before stream. At max maybe 6 hours, so the delay predicate would say something like that. This would get sent to every lava lamp regardless of its setting, but it would only care about the one. There are other ways of implementing this, like only sending the maximum of 6 hours and then having the lava lamp do the math on-board for when stream is and turning itself on in advance. Though i still think its possible to do without exposing anything.