Observability in the real world: Improve systems w/iterative approach
summary
Claim: deploy Datadog agents and a cluster agent to discover and instrument dozens of services without code changes using eBPF, Admission Controller, and OS runtime hooks; demo shows service discovery and working distributed traces within about five minutes and automatic generation of service topology and SLO-ready metrics. Workflows include live search, error tracking, RUM and watchdog summaries; Starbucks case shows starting SLOs from HTTP 200/500 logs and a 98% SLO baseline that sparks tenant collaboration.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 1:15Challenges in monitoring
- 4:10Observability approach
- 5:23Service instrumentation
- 11:01Non-kubernetes setup
- 12:53Metrics and traces
- 15:43Correlation and insights
- 17:10AI-driven monitoring
- 18:57Advanced monitoring
- 19:55Live demo: E-commerce app
- 29:06Demo: Non-k8s setup
- 31:21Demo: APM portal analysis
- 35:32Starbucks case study
- 39:31Future improvements
- 45:44Conclusion
related talks
38mDemocratizing Observability
33mEffective observability in microservice architectures
29mObservability; a gentle introduction
36mObservability from 0 to 100
49mImproving Observability of Distributed Systems
38mOff the Beaten Path: Moving Observability Focus from Your Service
01h 02mAdding Observability to Distributed Systems
38mAWS Observability as Code: Leveraging Datadog