Serverless Days!

Click here to see the full agenda

I had the amazing opportunity to attend ServerLess Days. If you are working with serverless technology and would like to meet other active developers and technical leaders from across the Australia & New Zealand region, while hearing from some world-class speakers, I highly recommend attending this event!

Here is a quick recap of my learnings!!

Think, Architect, and Build Serverless Applications as Set Pieces

Speaker: Sheen Brisals

  • Accumulated architecture HIDES complexity
  • Distributed architecture SHOWS the complexity
  • Problem Decomposition
    • Utilise serverless characteristics
      • Event-driven architecture
      • Using managed services
      • Infrastructure-as-code ♥
    • Employ PROVEN patterns
    • SUSTAIN the architecture, application and processes 🌳
      • ⬇ sustainability: Waterfall development
      • ⬆ sustainability:
        • Serverless ways of thinking
        • Evolving architecture
        • Minimum Viable Product (MVP 🌟)
        • Modular, extendable and observable product
      • Follow lean principles
      • Improve Value Stream
      • Agile & Pragmatic
      • Start SMALL and scale!
      • Evolve in ITERATIONS
      • Automate EVERYTHING 😎
      • DevOps Mindset
      • Refactor Continuously 💖
  • Growing talent in the team 🌻
    • Enablement within the team 🌼
    • Celebrating wins
    • Instill ownership

Whether Serverless or Not, It’s all about how you drive your project!

Speaker: Nelly Sattari

  • Keywords: Accountable, Agile, Healthy and Focused
  • The art of impossible
    • Coach - Helping individuals
    • Connector - Building a healthy team
    • Conductor - Driving projects
    • Compass - Setting the direction
  • Don’t forget to connect with stakeholders
  • Atlas Atlassian product that is a teamwork directory. For example, every Monday we can review how the team is doing!
  • Making it clear what everyone’s roles and responsibilities are
  • Building an Agile Team
    • Fixed: People, Team and Scope
    • Epic ▶ Milestone
      • Incremental delivery
      • Breakdown MVP
      • Carry out within 1 to 2 sprints
    • Importance of Sprint Goals ⚽
      • Achieving Sprint Goals is MORE important than shipping all the tickets
      • Each milestone could be a goal?
      • Goal setting meeting to review if the team met their goals in the previous sprint
        • Teams can answer YES, PARTIALLY or NO
        • If the sprint goal hasn’t been met, have a meeting to identify why the goals were not met
        • Other Agile Rituals include Sprint Checkpoint and Capacity Planning
  • 💟 DATA DRIVEN LEADERS 💟
    • Identifying work that is carried over multiple sprints, what are the dependencies? Why was this work not completed within the sprint?
    • Measuring velocity
    • Measure everything, no “gut feelings”
  • Connection
    • Social events
    • Celebrating milestones
    • Recognition
    • Beware of burn-outs!
  • Collaboration
    • Providing training
    • Mentoring and pairing
  • Data-Driven Engineer who recognises importance of Agile rituals
  • More team members can cause overhead
    • Adding more team members doesn’t mean the team will have higher velocity
  • Developing self-healing / stateless / idempotent systems
  • Working on serverless and real time observability
  • Summary
    • Do the simple thing first
    • Small team with fast feedback loop - showcase often
    • Identify risk early, shift-left and spike
    • Continuously measure performance and stress test
    • Isolate context boundaries (lambdas)
    • Solution must prove itself correct

Using serverless functions for real-time observability

Speaker: Liz Fong-Jones

  • Most realtime bulk workloads benefit
    • Move state from local machines onto object storage
    • Shard list of objects into work units
    • Parallelise object processing
    • Reduce results outside Lambda afterwards
  • Beware the dragons! 🐉
    • Avoid latency-insensitive batch workloads (cost)
    • Avoid tiny workloads (set-up latency)
    • Check cloud provider limits, state your intentions (capacity planning)
    • Test cross-compilation including profiling to avoid performance issues with switching architectures