Temporary host unreachable between containers in same docker network


we sometimes see the issue, that docker containers are not able to talk to each other, although they are in the same docker network, on the same host. It’s only a single RHEL7 server running docker-ce. Sometimes, after a restart, I see that the one container cannot reach the other. When I open a terminal on one container and try to ping the other, I see a host unreachable error. It shows that both containers are in the same subnet, and docker configuration shows the containers are both in the same docker network. Usually this gets fixed by a restart of the docker service. Maybe after that 2 other containers cannot talk to each other any more and I have to repeat this multiple times until all is fine.
I see this behaviour since years now, and multiple updates of the docker version did not fix it. I have no idea where to look any more, it just sometimes does this, I cannot reproduce it.

Does anyone have an idea what might cause this? How can I analyse this any further?

Currently: docker-ce-20.10.12-3.el7.x86_64 on RHEL7, EC2 instance on AWS