The road to world-class observability
summary
Describes a clear, actionable path: prioritize baseline monitoring, stabilize noisy alerts, add technical and business domain metrics, then automate SLIs/SLOs. Claims a 97% reduction in mean time to resolution after focusing on detection and alert quality. Uses concrete tools and practices: Terraform-generated New Relic dashboards as building blocks, PagerDuty automation actions for runbook-triggered fixes, DORA metrics for prioritization, and recommends 95th/99th percentile thresholds and anomaly-based (stddev) defaults.
AI-written summary · about
chapters (9)
- 0:00Introduction and journey to observability
- 3:38Maturity stages of monitoring systems
- 7:12Lowering incident resolution time by 97%
- 12:10Four golden signals for service health
- 18:50Standardizing monitoring with Terraform
- 23:07Stabilizing alerts and reducing noise
- 26:43Domain metrics for business and technical insights
- 30:40High-level monitoring dashboards and status pages
- 33:52Automating SLAs and continuous improvement
related talks
39mActionable Observability
29mObservability; a gentle introduction
31mNo-Nonsense Observability Improvement
38mDemocratizing Observability
23mMastering Monitoring & Alerting: Key to Seamless Performance
52mHolistic Monitoring
58mObserving Applications in Production
47mObservability - Holistic Quality across Software Systems