Building Observability for 99% Developers
summary
Argues that most teams—“99% developers”—lack time and expertise for heavy instrumentation and that observability must work with messy, aging stacks. Describes a concrete Akita solution: ebpf-based passive traffic listening to capture API calls with no code or proxy changes, plus automatic traffic modeling that infers API structure, field invariants and performance baselines and raises alerts when endpoints or schemas change, so teams spot regressions without writing dashboards or spans.
AI-written summary · about
chapters (18)
- 0:00Intro
- 0:58Building observability for 99% developers
- 2:00Hi, I'm Jean and I work on observability
- 4:46First, there's the problem of sheer code volume...
- 6:02Then there's the rise of microservices
- 6:42Then there's the rise of APIs
- 8:41Software development has changed
- 9:35What the Software Development Life Cycle often looks like
- 11:59Software practices that have gone out the window
- 12:32The observability people have it right
- 16:47It turns out there are a lot of developers outside of the ones we hear from
- 17:43And they look different from who we're hearing from
- 19:41The state of observability today
- 21:54Our conversations about software development assume planned gardens
- 23:12What embracing the evolving software rainforest means
- 24:28Solution Part One: eBPF-based passive traffic listening
- 25:34Solution Part Two: Automatic Traffic Modeling
- 26:18Understand your system behavior with Akita
related talks
40mBuilding Observability for 99% Developers
29mObservability for Developers: How to Get From Here to There
31mObservability for Developers: How to Get There
44mObservability: Superpowers for Developers
33mWhat Developers Should Know About Observability
23mDeveloper Observability - the Forth Pillar
18mNo-Frills Observability For The No-Nonsense...
48mAPI Observability by Design: Building Systems That Tell Their Own Stories