Intelligent network performance metrics
presentationposted on 2019-05-15, 00:26 authored by Richard Procter
REANNZ built TSX, a system that gives us accurate, long-term network performance metrics by abstracting away the continual changes to our network. It builds upon ideas found in the opensource Prometheus monitoring system.
In 2018, we coupled TSX to Prometheus in order to glean significant needles from our data haystack of sixty thousand metrics in near real-time. This has, for example, allowed us to proactively identify and replace ailing equipment before it fails, increasing the reliability of our network.
We will show TSX in action, how it has untangled our raw data store, and is now providing rich graphical reporting in both near-real-time and from our collection of over 5 years and nearly 2TB worth of recorded data.
ABOUT THE AUTHOR(S)
Richard Procter joined REANNZ in 2016 as a senior software engineer with a fifteen year background in commercial programming that spans telco-grade switching systems, open source network stacks, automated network design, and award-winning application development. At REANNZ he is responsible for network monitoring, network statistics and software practices.