From emailing irl: > The best I can come up with (for now) for automated ssh key selection: > > aws cloudformation deploy --region us-east-1 --stack-name `whoami`-onionperf-dev --parameter-overrides myKeyPair="$(./identify_user.sh)" --template-file onionperf-dev.yml > > ...where identify_user.sh is: > #!/bin/bash > aws ec2 describe-key-pairs | jq -r '.KeyPairs[].KeyName' | grep `aws iam get-user | jq -r .User.UserName` > > This assumes that: > 1. The username currently logged in on the machine is part of the key pair name > 2. There is only one key that will match a username. It would be nice to use more than one key, however AWS does NOT support multiple keypairs during stack creation/deployment > 3. The developer has jq installed > > I dislike this solution as it is very fragile but not sure where to go from here....thoughts? Other stuff already depends on jq. You can't easily do aws without it. Instead of fuzzy matching though, can we have static mappings in a table somewhere of username to key name? We probably only have like 3 or 4 users maximum.