I didn't see the math explicitly worked out in the thread so I worked it out myself. Apologies if this has already been posted elsewhere.
The proc rate of Replenish (15%) and the structure of Rejuvenation (4 ticks in 12 seconds) means that a given Rejuvenation has just under a 48% chance to proc one or more Replenish effects. The chances of procing a given number of Replenish effects off a single Rejuvenation are:
which means that although a given single Rejuvenation has just under a 48% chance to proc one or more Replenish effects, each Rejuvenation gives 60% of the Replenish effect on average. This means that in the long term (yes, I know fights don't last forever) Replenish can be thought of as:
Your Rejuvenation spell also restores 0.5 Energy, 0.25 Rage, 0.1% Mana or 0.5 Runic Power per second.
I don't know too much about energy/rage/RP generation rates, but 10 mana per second (MPS) on 10k mana seems like a nice buff to deliver with Rank 1 Rejuvenations that cost 1.6 MPS in ToL. (MPS will probably replace MP5 as a stat thanks to the new beta regen change.)
Now, if Rank 1 Rejuvenation works, then using anything higher for Replenish purposes is silly. If Rank 1 Rejuvenation doesn't work, Replenish gets bad in a hurry. It seems a little inelegant to me to create a high-level raiding mechanic that relies on 1.0s GCD Rank 1 Rejuvenations, and Blizz has a history of deprecating the use of very low-rank spells in raiding environments. For this reason, I don't think we've seen the final version of Replenish quite yet.
I can't comment on what I think this final version will be, as I'm not experienced with the mechanics and balancing of other party-regen abilities.
Note: Kortar informed me that I made this way too complex, and I did. The expected number of Replenish procs per Rejuvenate is just the number of Rejuve ticks times the proc rate, so 4*0.15 = 0.6.