Upgrade to Pro — share decks privately, control downloads, hide ads and more …

A Non-Technical Kubernetes Talk

569f10721398d92f5033097ac6d9132c?s=47 Tim Hockin
February 23, 2019

A Non-Technical Kubernetes Talk

Some of the squishier parts of Kubernetes.

KubeDay, Feb 2019

569f10721398d92f5033097ac6d9132c?s=128

Tim Hockin

February 23, 2019
Tweet

Transcript

  1. A Non-Technical Kubernetes Talk KubeDay Feb 2019 Tim Hockin <thockin@google.com>

    Principal SW Eng, Google thockin (c) Google LLC
  2. Who is this guy?

  3. The first Kubernetes developer meetup* Once upon a time, we

    all fit into one room * not really, but it did feel that way
  4. Community is at the heart of our success When we

    all fit in one room, it was easy - we all knew each other Now: one of the largest OSS projects in history • Contribs on almost every continent and timezone • Hundreds of sub-projects and thousands of spin-offs Large communities are VERY different from small ones Community, community, community
  5. Harder to trust people you don’t know Harder to communicate

    Harder to reach agreement Easier to fall into “us - them” Easier to forget something Easier to assume malice I have PERSONALLY fallen victim to all of these Harder and easier, and not in a good way
  6. None
  7. Commit to making our community stronger Look for opportunities to

    fix cracks, rather than make them Give people the benefit of the doubt Over-communicate Resist “us-them” language Help someone else do the work that you could have done yourself Call to action #1
  8. None
  9. We’re at the hardest phase of a project - and

    have been for a while (and will be) All the easy stuff is done, most of the obvious stuff has been brought up We have REAL users and customers: • Fortune 10 companies • Banks • Mega-stores • Websites that serve measurable percentages of the internet’s total traffic Day-to-day
  10. None
  11. We’re also not as robust as we need to be

    Governance and processes are evolving Tools are getting better (think back just 1 year!!) but are not done Testing is not where we need it to be We have a thousand features that intersect in subtle ways Growing up
  12. None
  13. Chop wood and carry water Part of our project’s ethos

    since forever Before you do a feature: fix a bug, improve a test, help with tooling Be defensive of our users’ trust Have patience - see call to action #1 Call to action #2
  14. Excitement around k8s comes from people seeing where it can

    take them Meet them where they are but also show them a future As systems mature, it’s VERY hard to keep that property We can not stand still - users’ needs are evolving We must not stagnate! Evolution
  15. None
  16. Skate to where the puck will be Look to the

    future Don’t design to hypotheticals, but do try to get ahead of the curve Give users what they need, not just what they ask for Call to action #3
  17. None