Can this lead to any problem, suboptimal performance etc. , if the RG0 is on the different node then the rest of the redundancy groups? RG0 has no reth interfaces, all the reths are in another group (RG1) - so there is no Z-traffic / this is purely active/passive... RG0 has only lo0. No tunnels are terminated on lo0 or anything fancy… But SRX has UTM/IDP features configured.
There is no problem or performance issue associated with the staggerred state you have explained.
RG0 -RE - Active on one node
RG1+ - PFE - Active on another node
The only difference in the scenario you describe versus all RGs being active on the same node, is that the RE-Active PFE communication will be over the physical control link connecting the two nodes. Since a stable control link is a pre-requisite for a cluster, there are no issues with the cluster being in this state.