The cloud moved your monitoring cheese
This document discusses the history and evolution of monitoring from infrastructure to applications and the impact of cloud computing. It provides an overview of infrastructure monitoring in the 1980s-1990s focused on servers. Application monitoring emerged focusing on application mapping and transaction tracing. The rise of cloud computing disrupted traditional monitoring as applications moved to platforms where the underlying infrastructure was abstracted. The document emphasizes that monitoring in the cloud needs to prioritize end-user experience through real user monitoring and synthetic checks. It provides an example case study of a company that implemented browser monitoring, synthetic checks, and software analytics on a PaaS platform to gain insights.
The cloud moved your monitoring cheese
- 1. The Cloud Moved Your Monitoring Cheese Ken Ahrens
- 2. Ken Ahrens Background Most of my career with monitoring vendors New Relic, CA, and startup ILC (Rockwell Collins) Also worked on Developer/QA efficiency API Virtualization and Testing Lived in Atlanta for 19 years Proud papa of 5 children
- 3. Agenda History of Monitoring Infrastructure Monitoring Application Monitoring Along comes THE CLOUD There went your cheese Now what!?
- 4. History of Monitoring Tomorrow Business Today Users 2000s Apps 80s - 90s Time Server and Networks Value Sources: Back to the Future, New Relic
- 5. Infrastructure Monitoring FCAPS (ISO Standard) Fault Configuration Accounting Performance Security Servers are your friends Source: Infoworld
- 6. Infrastructure Monitoring Vendors Source: SNMP Center
- 7. Application Monitoring Manually ARM your code Top-down approach Application Components/Map Transaction Traces Timers, Errors DevOps Identify impact of a release Cross-Infrastructure
- 8. Real User Monitoring Source: Could They Do It? OReilly Complete Web Monitoring
- 9. That Cloud Thing Everyone Is Talking About Source: The Onion
- 10. Is cloud primetime yet? Source: The Enterprisers Project Source: SRG Research
- 12. Monitoring Priority in Cloud IaaS High CPU is good, otherwise youre wasting $$ PaaS APM is a must-have you cannot see the server! SaaS May still be able to collect metrics via API
- 13. End User Experience On average 87% of the response time is in the Browser Your monitoring strategy should start with user experience Source: The Performance Golden Rule
- 14. How do you do that? Real end-user monitoring Instrument website Page timings and errors Classify by browser & geo Track AJAX calls Remote session traces Languages: JavaScript Synthetic monitoring Availability 24/7 checks against the site HTTP Response Codes Performance Baseline for trending analysis From multiple locations
- 15. Software Analytics Consolidate events from monitoring agents Enrich with business context Make data-driven decisions Focus on real-time
- 17. PaaS: Built on Shopify