Using the Vamp CLI

Vamp’s command line interface (CLI) can be used to perform basic actions against the Vamp API. The CLI was primarily developed to work in continuous delivery situations. In these setups, the CLI takes care of automating (canary) releasing new artifacts to Vamp deployments and clusters.

See also


Check the download page for details on how to install the Vamp CLI on your platform.


After installation, set Vamp’s host location. This location can be specified as a command line option (--host)

vamp list breeds --host=

…or via the environment variable VAMP_HOST

export VAMP_HOST=

Simple commands

The basic commands of the CLI, like list, allow you to do exactly what you would expect:

> vamp list breeds
NAME                     DEPLOYABLE
catalog                  docker://zutherb/catalog-frontend
checkout                 docker://zutherb/monolithic-shop
product                  docker://zutherb/product-service
navigation               docker://magneticio/navigation-service:latest
cart                     docker://zutherb/cart-service
redis                    docker://redis:latest
mongodb                  docker://mongo:latest
monarch_front:0.1        docker://magneticio/monarch:0.1
monarch_front:0.2        docker://magneticio/monarch:0.2
monarch_backend:0.3      docker://magneticio/monarch:0.3
> vamp list deployments
NAME                                    CLUSTERS
1272c91b-ba29-4ad1-8d09-33cbaa8f6ac2    frontend, backend

CI and chaining

In more complex continuous integration situations you can use the CLI with the --stdin flag to chain a bunch of commands together. You could for instance:

  • get an “old” version of a breed with inspect
  • generate a new breed based on the previous one, while inserting a new deployable
  • create the breed in the backend
vamp inspect breed frontend:${OLD} | \
vamp generate breed --deployable mycompany/frontend:${NEW} frontend:${NEW} --stdin | \
vamp create breed --stdin

Once you have the new breed stored, you can insert it into a running deployment at the right position, i.e:

  • get a blueprint from a running deployment with inspect and --as_blueprint
  • generate a new blueprint with generate while inserting a new breed
  • deploying the result with deploy
vamp inspect deployment $DEPLOYMENT --as_blueprint | \
vamp generate blueprint --cluster frontend --breed frontend:${NEW} --stdin | \
vamp deploy --deployment $DEPLOYMENT --stdin

See also