The single best method to this is complex political and economic concerns. In short, without there being negatives to war there is no reason not to war.
I picture the following scenario as the ideal way that avoids this: You have groups 1-6 here. Group 1 wants to fight a war. Group 2 is nearby, and a potential target. However, Group 2 shelters Group 3 from Group 1, and Group 3 provides a key resource in trade that Group 1 really needs and is not getting elsewhere. Group 4 is also a potential target with a border. It has relations with Groups 5 and 6, but those are only trade based and attacking would not lead to issues with supply or create border threats toward them.
Group 1 decides to fight Group 4. They fight for a while, but once a critical point in the fighting is reached Groups 2, 5, and 6 all step in for their own interests. After a while of combat, some of the group maybe have a little territory gained or lost, but they all need some time to rebuild stocks and prepare more weapons, armors, etc. So the fighting winds down for a while.
The keys to this are: Interests of various groups not being best suited by mega-grouping for the wins, complex economic factors and stocking up goods not being too easy, and benefits to the economic production when things are relatively peaceful (an easy example is that craft focused people left as war starts up need to continue work, but have bandits and monsters interfere and cannot do nearly as much as normal).
How feasible all that is, and how forgiving players are to the limitations thus created, is a giant question at this point.