Skip to content


Descriptor files are placed in a k8s directory in the root of your project file. This directory contains all the yaml files used to describe the Kubernetes deployment tasks needed to run this service. Target specific files are handled by using a -<target> "suffix", i.e. ingress-prod.yaml.

Files with a .yaml suffix will be applied to the Kubernetes cluster.

If both a common file and a target-specific file with the same "basename", i.e. service.yaml and service-local.yaml, only the target-specific file will be applied for a target and the common file will be applied to all other targets.

Files with a .sh suffix will be run on the machine executing the deploy command. This can be useful to setup secrets/configurations, mostly for local use. Note that only .sh files matching the target using the rules in the above paragraph will be executed, with the difference that no common files will be executed.

All other files in k8s will be ignored by the deploy command.

Available variables

The following variables are possible to use for substitution in the descriptor files:

Parameter Description
IMAGE The full image name (registry/name:tag)
COMMIT The commit SHA (3b701067e6a6943c773b9dc183fcc39cd31a2ff0)
TIMESTAMP The current time (2022-02-22T09:16:01+01:00)


$ cd projecct
$ tree
└── k8s
    ├── deploy.yaml
    ├── ingress-local.yaml
    ├── ingress-prod.yaml
    ├── service.yaml
    ├── service-local.yaml

Given the structure above:

deploy local

Will apply deploy.yaml, ingress-local.yaml and service-local.yaml and execute

deploy staging

Will apply deploy.yaml and service.yaml.

deploy prod

Will apply deploy.yaml, ingress-prod.yaml and service.yaml.