Day 2 Operations Best Practices with Apache Mesos
summary
Shows a working pipeline that polls the Mesos metrics API every 10 seconds from an 8‑node test cluster (7 private agents, 1 public agent), extracts master/agent/framework/executor counters and gauges with dimension tags, and sends them to SignalFx for dashboards. Implements the Mesos metrics module/collector that exposes UDP env vars, demonstrates spotting hosts at 100% CPU, interprets p90 vs p50, warns against high‑cardinality dimensions, and provides runnable scripts and alerts.
AI-written summary · about
chapters (27)
- 0:00Intro
- 1:11Agenda
- 1:52Continuously Connected World
- 2:41App Transformation
- 3:37Data Intensive
- 4:30Common Challenges
- 4:49Scalable Capacity
- 5:17Dynamic Architecture
- 6:12Load Balancing
- 7:19Metric Categories
- 8:02Mesos Architecture
- 8:38Metric Sources
- 9:02Master Metrics
- 14:01Marathon Metrics
- 14:13Master Basic Alerts
- 14:35Container Level Metrics
- 15:18Mesos Metrics Module
- 16:11Metrics API Architecture
- 18:27Metrics API Response
- 18:40Datapoint
- 19:03Metric Types
- 20:20Metrics vs. Dimensions
- 20:37Metric Time Series
- 21:14DC/OS Metrics API Usage • Get authentication token
- 22:00Sending Metrics: Do...
- 23:06Monitoring
- 27:42Key Takeaways
related talks
18mWhat if Mesos Metrics Collection was a Snap?
22mEnd-2-End Monitoring and Troubleshooting a Real-Time Data Pipeline
29m“If you Don’t Monitor your Infrastructure, you Don’t Own it!” Regain Control Thanks to Prometheus
34mDay 2 Operations : Make Friends with Your Ceph Cluster
41mMetrics Applied
52mIs your JVM app flying blind? Unmask hidden issues with Observability superpowers!
48mMetrics All the Way: Data Driven Operations
28mPrometheus Operators in the Wild