Proactive Monitoring @Twitter

Most systems provide statistics for tracking usage, performance, and failures. However, some do not fit into pre-existing monitoring infrastructure, or provide no metrics at all. This talk discusses a proactive approach to monitoring services at Twitter. Using case studies from multiple services at Twitter, we illustrate the value of a holistic approach to monitoring from design review up to tuning alerts in production.

Joe is one of the founding members of Twitter’s Aurora/Mesos SRE team, and collaborated to build a highly available tiered fleet at Twitter. He has contributed to everything from the Puppet infrastructure that orchestrates Twitter’s Mesos cluster all the way up to automation of maintenance, deploys, and upgrades. Before that, he was part of Google's Internal Technology Residency Program in Mountain View, Boston, and Zurich. He holds a B.S. in Computer Science from Chapman University.