Your Toolchain is Production: The Case for Observability
summary
Argument states the developer toolchain is a production system and needs the same observability as customer-facing services: logs, metrics, traces to diagnose why an hour‑long build can take 55 minutes to fail or why flaky tests force hours of retries. Adds three more signals as TEMPLE (exceptions, profiles, events) and shows a concrete product, Develocity Providence Governor, issuing signed attestations for continuous GRC. Cites ~50–60 vulnerabilities published daily and demonstrates provenance gaps that prevent tracing artifacts back to build agents.
AI-written summary · about
chapters (9)
- 0:00Defining the developer toolchain as production
- 1:16Common scenarios of toolchain failure
- 3:00Business impact of toolchain issues
- 5:39Key requirements for toolchain stability
- 7:42The three pillars of observability: logs, metrics, traces
- 11:46Introducing TEMPLE: exceptions, profiles, and events
- 13:21Improving security and GRC practices
- 15:20Benefits for engineers and managers
- 16:40Strategies for getting started
related talks
13mForget Production: Observability a Non-Production Necessity
13mForget Production: Observability a Non-Production Necessity
27mRethinking Observability for Off-the-Shelf Tools
44mObservability: Superpowers for Developers
23mA peek into observability from a tester's lens
20mKeynote: Prometheus, Jaeger, and Envoy: The Observability for Today's Enterprises
26mObservability and Its Role in the Software Life Cycle
33mUnlock Embedded Excellence: Reliability Secrets Revealed!