Leveling up your monitoring

In a welcome change of pace, people are paying more attention than ever before to what, how, and when they monitor their systems and applications. In this talk, we'll journey from level 0 (peasant) to … well, something much more powerful*, able to slay incident-dragons wherever they may appear. We'll be looking at concepts and practices more than specific tools, so there'll be something for everyone regardless of their background.

* Monitoring, like characters in a classic RPG, knows no level cap.

Transcript

1. Leveling Up Your Monitoring 12/12/12 Triangle DevOps - Ben Scofield - @bscofield
3. 0 Customer alerts
5. shouldn’t we do ... something?
6. LEVEL UP
7. 1 Site monitoring
9. what about user flows?
10. +XP
11. what about all of our edge cases?
12. LEVEL UP
13. 2 Exception alerts
15. ... my mailbox is full.
16. +XP
17. what if the database blows up?
18. LEVEL UP
19. 3 System alerts
21. how do these incidents affect the business?
22. LEVEL UP
23. 4 Business metrics
25. let’s measure all the things!
26. +XP
27. are we measuring too many of the things?
28. +XP
29. do we need to watch all of the things all of the time?
30. LEVEL UP
31. 5 Anomaly alerts
33. welcome to the company! you’re on call
34. LEVEL UP
35. 6 Actionable alerts
36. couldn’t a monkey do this?
37. LEVEL UP
38. 7 Automation
39. it doesn’t stop
40. Thanks! 12/12/12 Triangle DevOps - Ben Scofield - @bscofield