The matching algorhytm for this rules did not check if the matched cluster did actually belong to the node.
For example if you had the following clusters
ClusterA: Host1, Host2 ClusterB: Host3, Host4And these "Clustered services for overlapping cluster" rules
Rule 1) Service1 -> ClusterA Rule 2) Service1 -> ClusterB
If Host3 has a Service1, the rule number one would match and assign this service to ClusterA, ignoring the fact that this host is actually a part of ClusterB.
This has been fixed.