Open codificat opened 2 years ago
/priority important-soon
/triage accepted as this was the intention from the tech talk where this was proposed
I think this is not planned for the current sprint?! /priority backlog
/sig devsecops
Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale
.
Stale issues rot after an additional 30d of inactivity and eventually close.
If this issue is safe to close now please do so with /close
.
/lifecycle stale
/remove-lifecycle stale /lifecycle frozen
The way you can do this with Openshift BuildConfig is pretty elegant: Build of parent triggered -> ImageStream X change -> BuildConfig with trigger on X trigger a build -> chain and repeat ad nauseam.
We might want to take a lookt at Tekton Trigger to see if it's feasible to hook them up with ImageStream ; or what else we can use to detect new images tags / versions.
/remove-lifecycle frozen /priority backlog
Is your feature request related to a problem? Please describe.
Every time an update is released is made on a package that is part of a base image, we have to wait for the change to be included in the whole chain of images that contain that package.
For example, an update can take this path: s2i-minimal -> s2i-lab -> s2i-custom-notebook, and only then we can use the actual "end user focused" image containing the update.
Currently, each of these updates is handled manually.
Many projects depend on this type of chains of base images.
High-level Goals
A possibility could be a new kebechet manager that works similar to the package updates but for container images.
With that in place, updates can be triggered manually by opening issues for Kebechet. Later, the creation of these issues can be automated following the image cascade path.
Describe alternatives you've considered
Keep it manual.
Additional context
Discussed in a community tech talk on 2021-11-18 (recording).
aicoe-ci pipelines can handle sibling releases, but this is parent/child dependency.
Acceptance Criteria