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

Where's the value in Cloud Native

Armakuni
September 04, 2018

Where's the value in Cloud Native

Whether you're a developer or a manager we will give you a framework to decide which cloud native practices you should adopt today. We explain the practicalities of implementing three high value cloud native practices, both from an implementation perspective, and the perspective of someone having to sell them to the wider organisation.

Armakuni

September 04, 2018
Tweet

More Decks by Armakuni

Other Decks in Programming

Transcript

  1. Where's the value in Cloud Native High value cloud native

    practices, and how to sell them to your boss
  2. Who are we Billie Thompson Cloud Native Consultant at Armakuni

    @PurpleBooth Zenon Hannick COO at Armakuni @ZenonHannick
  3. Key Values: Raw • Lead Time • Process Time •

    Complete & Accurate • Number of Dependant Systems
  4. Key Values: Observed • Work that falls outside VSM (Unplanned

    Work) • Variance in PT • Variance in LT • Common adjectives ◦ “Confusing” ◦ “Fast” ◦ “Boring”
  5. Manager • The (difficult) conversation • Shared understanding & agreement

    • Identifies best areas to focus effort • Shows flow of value! • Numbers & Data • Allows you to track progress • Can set future desired state
  6. Objection! • I can’t get everyone together ◦ Gather the

    statistics one to one • I don’t know all the times ◦ Find someone who does ◦ Educated guess ◦ Dig it out of a tool ▪ VCS ▪ Deployment Tool ▪ Ticketing System
  7. The Symptoms • Merging processes • Test Deployment processes ◦

    High PT • Deployment processes ◦ High LT ◦ High PT ◦ Low C&A ◦ High Variance of LT
  8. Practitioner • Building software vs Building a system • We

    share the codebase • My laptop won’t fit in a server rack • Oops I forgot to… ◦ Run the tests ◦ Check the migrations ◦ Run the security tests ◦ etc
  9. Manager • Visualises flow of code (value) • Encourages constant

    flow • Speed • Smaller batch sizes • Reduces risk • Fully auditable trail
  10. Objection! • I need to press a button ◦ Illusion

    of confidence in manual step ◦ Lack of trust in automated testing • I need to review changes • Governance ◦ Pipeline automatically provides audit trail ◦ Can automate governance checks as part of the pipeline • Security ◦ Add to pipeline
  11. The Symptoms • Deployment Process ◦ High LT ◦ High

    PT ◦ Poor C&A • Unplanned Work ◦ That happens at 3am ◦ Easily recoverable errors ◦ Scaling • Large % of time & effort operating platform/infrastructure, (i.e. non-value adding work)
  12. Practitioner • No need to learn about servers • My

    code, my responsibility • Prototyping is easy • I no longer need to think about: ◦ Reliability ◦ Scaling ◦ Monitoring
  13. Manager • Focus on business value delivery • Abstracts complexity

    • Standardise across organisation • Software delivery practices for platforms • Turns platform team into a product team • Speed • Less things for people to worry about
  14. Objection! • I need to know everything that is going

    on! • I need to control the whole stack • We are very special unique snowflake
  15. The Symptoms • Development processes ◦ High variance ◦ Poor

    C&A • Common Adjectives ◦ “Confusing” ◦ “Complicated” • Unplanned work ◦ Misunderstanding system ◦ Wrong prioritization
  16. The nitty gritty Set some additional headers, with HTTP: •

    X-B3-TraceId • X-B3-ParentSpanId • X-B3-SpanId • X-B3-Sampled or X-B3-Flags (Or a bunch of other alternatives) Optionally an API call to the OpenTracing service.
  17. Practitioner • What calls what is no longer a mystery

    • Makes me think about users • O(n) calls become really obvious • I can prove it!
  18. Manager • Observability • Evidence of areas needing improvement •

    Enables association of user activity to system activity