Ok our setup is this AWS ALB -> 2x EC2 Dockerized nodes -> Nginx -> services.
For a brief period <5 minutes, the ALB could not contact the any services running on any of the nodes.
Resulting in 504 from the aws/elb2.0 (ALB)
No restart or redeploy was done. Then it magically came to life a few minutes later.
At the same time I could connect via the nodes io addresses.