Branch: master
Find file History
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
..
Failed to load latest commit information.
999-shipperctl-admin.adoc
README.md

README.md

Roadmap

This document contains the outline of our Roadmap brainstorm session. The idea is to transform this outline into different RFCs and discuss them as separate Pull Requests.

Communication Channels

Web Presence

  • Appealing web site/Github pages.
  • Documentation
    • Architecture
      • Use cases helps defining the release workflow.
      • Release workflow diagram helps outlining required tests.
    • Operations manual
      • How to install Shipper?
      • How to manage Shipper?
        • How do I include a cluster?
        • How do I cordon a cluster?
        • How do I drain a cordoned cluster?
    • User manual
      • How do I use Shipper?
      • Knowledge base of problems, solutions and possible and workarounds

Support and announcements

  • Slack
  • Announcement mailing list

Marketing

  • Blog posts
  • Meetups
  • Kubernetes community engagement strategy

Technical Improvements

  • Improve unit and end-to-end tests infrastructure.

    • Investigate and design a reliable testing process.
      • Get rid as much as possible of boilerplate code.
    • Implement the designed testing process.
  • Structured errors library.

  • Even conditions implementation in current code-base.

  • Revisit the current architecture.

    • Controllers design: how to structure "act" and "observe" operations to decrease unnecessary API calls.
  • Reporting efficiency metrics.

    • API calls vs. Releases.
    • What is the minimum number of API calls required to perform a successful rollout?
  • Revisit status reporting.

    • Should we consider Conditions as API?
    • Are outlook variables a good thing after all? Is there anything that we can't do with conditions (waitingForCommand, for example).

Technical Innovations

  • Rollout blocks.

  • Capacity fleet summary.

  • Improve end user release level information.

  • shipperctl

    • Cluster admin operations.
      • Facilitate complicated operations such as adding a new cluster into the Management Cluster.
      • Only complex commands; avoid implement wrappers things that a kubectl command would suffice.
    • User operations.
  • No-op rollouts to facilitate cluster draining in case of cluster maintenance and cordoning.

  • Investigate and design Shipper high-availability strategies.

  • Introduce a HTTP API to summarize Application data (both Management Cluster and Application Clusters' objects).

    • Is my application being rolled out?
    • What is the status of my current deployment process?
    • Which Pods are failing to initialize, per cluster?
  • Introduce a Web based UI to display summarized Application data.

  • Investigate and design support off-the-shelf charts.

  • Capacity based scheduling.

  • Investigate and design a more refined strategy language.

    • Consider per-cluster vanguard.
    • Consider having strategy sub-steps optional (for example, manage only installation and capacity, but not traffic shifting).