20170901_srecon17_1.6
20170901_srecon17_1.6
Monitoring design principles
0 - Agenda
- History
- Service Level Objectif
- Simple designs
1 - History
Monitoring:
- Action of Observing, checking, and compute a status
Evolution of Web monitoring
- 1995: Load webpage every 15 minutes
- 2000: Wathcing every webrequests
- 2015: Deep analysis of every transaction from a user perspective
Evolution of DB monitoring
- 1995: Synthetic queries to test perf
- 2000: Wathcing every queries
- 2015: Deep analysis
Monitoring is sophisticated
* Increased telemetry volume
* More valuable operational questions
* Increased organizational velocity
2 - SLO
Usually based on percentiles
3 - Best practices
- Monitoring outside of the Tech stack / blast radius
- Do not silo your data (For correlation)
- Alerts require documentation
+ Requires discipline
+ Human readable explanation
+ Business impact description
+ Remediation procedure
+ Escalation documentation