Control Theory for SRE
Talkat SREcon19 Europe/Middle East/Africa · on Wednesday, 2019-10-02· at 14:00· Track 1: The Liffey B
summary
Builds a working PID controller that autoscaling Kubernetes can use and shows what actually happens: a simulated vehicle uses location as the process variable and velocity as the control output, sampled every 1s and displayed every 5s, with an I window of ~10 minutes and clamped actuator values. A pure P term oscillates forever; adding I and D converges slowly and reduces overshoot. Implementation uses Prometheus for metrics, Go binary as the controller, Grafana dashboards, and code on GitLab; a live pod-based Kubernetes model is planned.
AI-written summary · about