Keep chain tip in memory for sync_chain_mmr endpoint

[sergerad]

Not for this PR, but it may be good to keep this info in memory somehow so that we could avoid going to the DB just to look up just to figure out which block number to use. Maybe this could be returned from State::latest_block_num() - e.g., as something like:

let chain_tip = self.state.latest_block_num(finality_level).await;

Let's create an issue for this.

Originally posted by @bobbinth in #1725 (comment)

        // Determine the effective tip based on the requested finality level.
        let effective_tip = match request.finality() {
            proto::rpc::Finality::Unspecified | proto::rpc::Finality::Committed => chain_tip,
            proto::rpc::Finality::Proven => self
                .state
                .db()
                .select_latest_proven_block_num()
                .await
                .map_err(SyncChainMmrError::DatabaseError)?,
        };

[sergerad]

@bobbinth @Mirko-von-Leipzig for this I am assuming we will be doing something like this in State

    /// The latest proven-in-sequence block number, updated by the proof scheduler.
    proven_tip: watch::Receiver<BlockNumber>,

Which the proof scheduler sends to.

Just wondering whether we are OK with implementing that before we get into #1853?