Monitoring Mesos with InfluxDB & the TICK Stack
summary
Complete monitoring pipeline based on the open-source TICK stack: Telegraf agents collect Mesos, Docker and ZooKeeper metrics via ~100 plugins, data lands in a single InfluxDB instance sized 8 vCPU 32GB (multi-tenant), default retention 90-day retention, and Grafana serves prebuilt dashboards. Kapacitor runs tick scripts, persists alerts back into InfluxDB, and supports user-defined functions (example: TensorFlow anomaly detection). Watchdog autoscaler writes events to InfluxDB; overall cluster utilization averages 25% global utilization with some teams hitting ~60%, enabling capacity and cost analysis.
AI-written summary · about
chapters (18)
- 0:00Intro
- 0:49InfluxDB
- 2:28Timeseries Data
- 3:26Timeseries Types
- 6:05Query Interface
- 7:00Discovery
- 8:59Querying
- 11:13percentiles
- 12:42TICK Agent
- 13:57Cronograph
- 16:29Capacitor
- 19:12Tomas
- 19:50Focus
- 21:12Mesos
- 23:57Stack Overview
- 26:28Capacitor Overview
- 33:06Screenshots
- 36:45Summary
related talks
38mInfrastructure Monitoring Basics with Telegraf, Grafana and InfluxDB
42mDistributed Monitoring: How to understand the Chaos
23mHow to Monitor Application Performance in a Container-Based World
35mReal-Time Monitoring with Grafana, StatsD and InfluxDB- 42mDistributed Monitoring: How to understand the Chaos
42mUnified Monitoring of Containers and Microservices [I]
50mMonitoring and metrology for containers
30mMethods & Tools For Time Series Data Science Problems With InfluxDB