Current recommendation for pipelines as code


#1

I want to define my (Kubernetes) pipelines in code and am not entirely clear on the current recommended best practice.

Is this reference currently the best guide to getting started with this? https://github.com/spinnaker/dcd-spec/blob/master/ALPHA-GETTING-STARTED.md


#2

Hi,

Managed pipeline templates are reasonably supported now (managed through http://github.com/spinnaker/roer), and several users are happily using managed pipeline templates (MPT). That being said, please be aware that we are currently reworking our pipeline templating story and most likely adding a new CLI surface to accomplish this. We are planning to add migration support for MPT/roer users, but there will be some work in migrating pipeline using the MPT to the new workflow (likely jsonnet as input to the new CLI surface).

Let me know if you have further questions.


#3

Hi,

At Waze, we use managed pipeline templates very extensively (though mostly for VMs, and only some k8s).

Check out the Blog post about it here.

We are also presenting in GCP Next 2018, specifically about the benefits of Spinnaker managed pipeline templates & infrastructure as code, alongside Netflix and Google regarding large scale deployments and Kayenta.

Tom


#4

Here is a small blog post hopefully clarifying between DCD/managed pipeline templates and spin: https://blog.spinnaker.io/spin-and-roer-managed-pipeline-templates-4fde2951c648. Let me know if you have further questions.