Real-Time Monitoring with Grafana, StatsD and InfluxDB
summary
Shows a working pipeline: apps emit StatsD metrics over UDP to InfluxDB and Grafana dashboards built from those events, with Ruby snippets for timing and incrementing and tags for dimensions (document.type). Highlights practical pitfalls: Telegraf turns dots to underscores, and batching/flush makes naive COUNT queries wrong so use SUM aggregation, uses StatsD over UDP and Grafana + InfluxDB as the visualization/storage pair. Gives concrete wins: a deploy-caused page increment dropped to zero, a flaky fraud service triggered a UX change to create proposals, and vehicle-refusal counts drove product choices.
AI-written summary · about
chapters (25)
- 0:00Intro
- 0:29Alpha and Beta Centauri
- 5:15Dont Tell
- 6:29About Artur
- 8:43Log
- 10:22What do we do
- 10:59StatsD
- 12:23Refonax
- 13:14Dashboard
- 14:24Ruby syntax
- 16:10Tags
- 16:23In action
- 16:38Building the chart
- 20:33Improving the chart
- 22:16Creating the chart
- 22:31The question
- 24:28The normal flow
- 24:40Calling StatsD
- 25:08Blind Spot
- 27:54Risk
- 29:00Datadriven usercentric
- 30:48Tools
- 32:40Resources
- 33:55Share
- 34:38Conclusion
related talks
38mInfrastructure Monitoring Basics with Telegraf, Grafana and InfluxDB
54mA Practical Guide to Dashboarding
27mEasy App Metrics
49mA Practical Guide to Dashboarding
42mDistributed Monitoring: How to understand the Chaos
59mGathering Metrics With StatsD And Graphite
46mHow many dashboards does your Grafana have?
20mMonitoring, And Visualizing your Node.JS server with Open Source tools