Load balancer uses wrong (none existend) IP address

I’ve got a cluster with 1 controller, 3 workers as in the screenshot below. On creating a load balancer an IP address which does not exist anywhere is inserted as 1 of the 3 load balancer ingress addresses. The second screenshot shows this. The workers have 34.245.89.174, 18.203.243.91 and 34.247.91.201, while the load balancer entry has 18.203.243.91, 34.247.91.201 and 3.248.222.169. 3.248.222.169 is the one which does not exist. It belongs to the Amazon Ireland region IP range, as all the others, so my suspicion it it some kind of left over, maybe cached, IP address, from previous used hosts. I created hosts, dumped them, created new one, and every time EC2 assigns a new and different IP number. My question is how do I get rid of it and force the load balancer to use the correct IP address.

I can now, at least partly, answer my own question. By looking at the API for the cluster I noticed that the wrong IP address 3.248.222.169 was in the API assigned to worker1-1-1. I deleted this node and created a new one, after that the problem was gone, though it’s still unclear how the wrong IP came into place in the first instance.