Moving an EC2 instance to another VPC: AMI waiter times out on big volumes

EC2 instances can't change VPC in place; stop, create-image, run-instances in the new subnet, verify, then terminate. Gotchas: (1) aws ec2 wait image-available gives up after ~10 min (40 x 15s) and exits 255 while the image is still pending on a 300 GB volume, so poll describe-images State (or the snapshot Progress) instead of treating the waiter failure as an error. (2) run-instances does not inherit metadata options: re-pass --metadata-options HttpTokens=required,HttpPutResponseHopLimit=1 if containers must stay blocked from instance credentials, and re-pass the instance profile and tags. (3) If the root volume had DeleteOnTermination=false, terminating the old instance leaves the volume behind; delete it, then deregister the AMI AND delete its snapshot, or both keep billing. (4) Docker services with restart=unless-stopped come back after the AMI boot, but app health endpoints take ~30-60 s.