The AWS Companion app contains useful additions to the AWS cli.
Go to https://github.com/opsidian/awsc/releases and download the appropriate binary for your system.
Rename it to awsc and move it somewhere on your PATH.
Run awsc to see the available commands and flags.
The command expects you to use AWS profiles.
You can create your profile in the ~/.aws/credentials file in the following format:
[my-company-dev]
region=eu-west-1
aws_access_key_id=AKIA...
aws_secret_access_key=J5oAF9...
If you need to assume a role with MFA then you should also add an extra profile in ~/.aws/config:
[profile my-company-dev-some-role]
source_profile = my-company-dev
mfa_serial = arn:aws:iam::123456789:mfa/your_username
role_arn = arn:aws:iam::123456789:role/SomeRole
Usage:
AWS_PROFILE=my-profile awsc auth
The command will create temporary credentials and save them under ~/.awsc with the given expiration time. The command generates three files:
- ~/.awsc/my-profile.json: the temporary credentials in JSON format
- ~/.awsc/my-profile.env: the credentials exported as environment variables, so you can source them from a bash script
- ~/.awsc/my-profile: a helper script which sources ~/.aws/my-profile.env and runs the given command. It also runs awsc auth to automatically reauthenticate if necessary.
I suggest to add ~/.awsc to your PATH in your bash profile:
export PATH=$PATH:$HOME/.awsc
Example:
$ AWS_PROFILE=my-company-dev awsc auth
MFA token: ******
$ my-company-dev aws s3 list-buckets
If you plan to use the helper script, then you have to run the awsc auth command only once per profile.
If you want to use your own script then include these two lines before you interact with AWS:
AWS_PROFILE=my-company-dev awsc auth
. $HOME/.awsc/my-company-dev.env
awsc autoscaling migrate <auto scaling group name> --region <AWS region>
The command will terminate all auto scaling group instances one-by-one. When an instance is terminated it waits for a new instance to be created and be in service.
If your auto scaling group has only one instance then this command might cause downtime.
If you use your autoscaling group with an ECS cluster you can tell the command to drain your ECS instances first.
awsc autoscaling migrate <auto scaling group name> -ecs-cluster <ECS cluster name> --region <AWS region>
You can set the minimum percent of instances kept healthy during the migration process with the --min-healthy-percent option (default is 50).