For our 4th installment of the DevOps Dojo hangouts, we corralled monitoring mavens from Atlassian, LogicMonitor, Netflix, and Metafor for an hour-long geek-out that got quite animated!

Did you know, for example, that the very notion of distinguishing application monitoring from infrastructure monitoring is controversial? I did not. Or ever thought about the effects of human bias when looking for anomalies in monitoring data? Personally, it never crossed my mind.

There’s lots for your brian to chew on in here. So grab a sandwich (assuming you don’t subsit on brain-food alone) and fire up the video!


0:00 – Welcome

1:40 – What’s the difference between application & infrastructure monitoring?
6:50 – Retroactively instrumenting infrastructure to be monitored
14:30 – Instrumenting application code to facilitate infrastructure monitoring
20:40 – Business metrics vs. operational metrics (pt. 1)
22:15 – Tools of the trade
27:45 – Chaos Monkey, and other members of the Netflix Simian Army
32:15 – Anomaly detection
37:40 – Dashboards and their dark side
44:35 – Human bias vs. automated objectivity in discerning meaningful anomalies
47:15 – Sharing is caring (monitoring as a means of getting devs & ops to collaborate)
58:00 – Why monitoring isn’t boring


Ps: For more tips on keeping it real with DevOps, check out

Alerts, anomalies & automated canaries at Dev...