Continuous Delivery

Skaffold offers several sub-commands for its workflows that make it quite flexible when integrating with CI/CD pipelines.

skaffold run

skaffold run is a single command for a one-off deployment. It includes all the following phases as it builds, tags, deploys and waits for the deployment to succeed if specified. We recommend skaffold run for a simple Continuous Delivery setup, where it is sufficient to have a single step that deploys from version control to a cluster. For more sophisticated Continuous Delivery pipelines, Skaffold offers building blocks that are described next:

  • healthcheck - wait for deployments to stabilize and succeed only if all deployments are successful
  • skaffold build - build, tag and push artifacts to a registry
  • skaffold deploy - deploy built artifacts to a cluster
  • skaffold render - export the transformed Kubernetes manifests for GitOps workflows

Waiting for Skaffold deployments using healthcheck


skaffold deploy optionally performs a healthcheck for resources of kind Deployment and waits for them to be stable. This feature can be very useful in Continuous Delivery pipelines to ensure that the deployed resources are healthy before proceeding with the next steps in the pipeline.

To determine if a Deployment resource is up and running, Skaffold relies on kubectl rollout status to obtain its status.

Waiting for deployments to stabilize
 - default:deployment/leeroy-app Waiting for rollout to finish: 0 of 1 updated replicas are available...
 - default:deployment/leeroy-web Waiting for rollout to finish: 0 of 1 updated replicas are available...
 - default:deployment/leeroy-web is ready. [1/2 deployment(s) still pending]
 - default:deployment/leeroy-app is ready.
Deployments stabilized in 2.168799605s

Configuring status check time for deploy healthcheck

You can also configure the time for deployments to stabilize with the statusCheckDeadlineSeconds config field in the skaffold.yaml.

For example, to configure deployments to stabilize within 5 minutes:

  statusCheckDeadlineSeconds: 300
    - k8s-*

With the --status-check flag, for each Deployment resource, skaffold deploy will wait for the time specified by progressDeadlineSeconds from the deployment configuration.

If the Deployment.spec.progressDeadlineSeconds is not set, Skaffold will either wait for

the time specified in the statusCheckDeadlineSeconds field of the deployment config stanza in the skaffold.yaml, or default to 10 minutes if this is not specified.

In the case that both statusCheckDeadlineSeconds and Deployment.spec.progressDeadlineSeconds are set, precedence is given to Deployment.spec.progressDeadline only if it is less than statusCheckDeadlineSeconds.

For example, the Deployment below with progressDeadlineSeconds set to 5 minutes,

apiVersion: apps/v1
kind: Deployment
  name: getting-started
  progressDeadlineSeconds: 300
      - name: cannot-run

if the skaffold.yaml overrides the deadline to make sure deployment stabilizes in a 60 seconds,

apiVersion: skaffold/v1
  statusCheckDeadlineSeconds: 60
    - k8s-*

Running skaffold deploy

skaffold deploy --status-check

will result in an error after waiting for 1 minute:

Tags used in deployment:
Starting deploy...
kubectl client version: 1.11+
kubectl version 1.12.0 or greater is recommended for use with Skaffold
 - deployment.apps/getting-started created
Waiting for deployments to stabilize
 - default:deployment/getting-started Waiting for rollout to finish: 0 of 1 updated replicas are available...
 - default:deployment/getting-started failed. Error: received Ctrl-C or deployments could not stabilize within 1m: kubectl rollout status command interrupted.
FATA[0006] 1/1 deployment(s) failed

skaffold build | skaffold deploy

skaffold build will build your project’s artifacts, and push the build images to the specified registry. If your project is already configured to run with Skaffold, skaffold build can be a very lightweight way of setting up builds for your CI pipeline. Passing the --file-output flag to Skaffold build will also write out your built artifacts in JSON format to a file on disk, which can then by passed to skaffold deploy later on. This is a great way of “committing” your artifacts when they have reached a state that you’re comfortable with, especially for projects with multiple artifacts for multiple services.

Example using the current git state as a unique file ID to “commit” build state:

Storing the build result in a commit specific JSON file:

export STATE=$(git rev-list -1 HEAD --abbrev-commit)
skaffold build --file-output build-$STATE.json

outputs the tag generation and cache output from Skaffold:

Generating tags...
Checking cache...
 - Found. Tagging

The content of the JSON file

cat build-$STATE.json

looks like:


We can then use this build result file to deploy with Skaffold:

skaffold deploy -a build-$STATE.json

and as we’d expect, we see a bit of deploy-related output from Skaffold:

Tags used in deployment:
 - ->
Starting deploy...
 - pod/getting-started configured

skaffold render


Skaffold also has another built-in command, skaffold render, that will perform builds on all artifacts in your project, template the newly built image tags into your Kubernetes deployment configuration files (based on your configured deployer), and instead of sending these through the deployment process, print out the final deployment artifacts. This allows you to snapshot your project’s builds, but also integrate those builds into your deployment configs to snapshot your deployment as well. This can be very useful when integrating with GitOps based workflows: these templated deployment configurations can be committed to a Git repository as a way to deploy using GitOps.

Example of running skaffold render to render Kubernetes manifests, then sending them directly to kubectl:

Running skaffold render --output render.txt && cat render.txt outputs:

apiVersion: v1
kind: Pod
  name: getting-started
  namespace: default
  - image:
    name: getting-started

We can then pipe this yaml to kubectl:

cat render.txt | kubectl apply -f -

which shows

pod/getting-started configured

Or, if we want to skip the file writing altogether:

skaffold render | kubectl apply -f -

gives us the one line output telling us the only thing we need to know:

pod/getting-started configured
Last modified February 22, 2021: Update (#5429) (6d62cf9)