Holistic Monitoring
summary
Argues for monitoring by function not tool, listing nine functions (infrastructure, APM, availability, RUM, logging, metrics, dashboards, SLIs/SLOs, alert orchestration) and shows a working pipeline using Prometheus/Zabbix for infra, New Relic synthetics and Selenium for external checks, Grafana for dashboards, logs and tracing, and PagerDuty alert orchestration to group incidents. Demonstrates the origin-based availability fallacy and why blackbox user-centric checks are required. Claims root cause is found in under ten minutes and a slow query drops from ~4s to 70ms.
AI-written summary · about
chapters (21)
- 0:00Start
- 1:31Problem Statement
- 2:03Problems Statement
- 2:29Perception
- 12:52Holistic Monitoring Strategy
- 13:15Functions of Monitoring
- 14:57Signal-to-Noise Ratio
- 16:00Infrastructure
- 16:55Application Performance Monitoring
- 17:52Availability
- 18:35Origin Based Availability Monitoring
- 21:46Monitoring Strategy
- 23:25Real User Monitoring
- 27:04Service Level Objective
- 28:48Alert Orchestration
- 30:38Metrics and Dashboards
- 37:28Customer Portal Monitoring Strategy
- 40:35Logging
- 42:41Summary
- 44:07How Do You Get the Notifications from Your Monitoring to Your Alert Orchestrator
- 48:20How Do We Feel about Automatically Restarting Services Based on Moderate Monitoring Telemetry
related talks
23mMastering Monitoring & Alerting: Key to Seamless Performance
20mUsing Monitoring to Speed Up Your DevOps Journey
40mDemystifying Observability and Monitoring
36mThe road to world-class observability
41mMetrics Applied
34mReliability Monitoring: Tips and Tools
56mFull-stack observability and application monitoring with
58mObserving Applications in Production