Autoscaling with HashiCorp Nomad

on YouTube ↗30m

" summary

Deploys a Nomad job with nomad autoscaler using Prometheus metrics, Grafana dashboards and Loki logs to make real scaling decisions. Policy uses a target of five concurrent connections, min/max of 1/20 and a 20s cooldown; demo shows counts drop from 3→1 on zero traffic and climb when 30 concurrent users are applied. Autoscaler also drives AWS Auto Scaling Groups to add clients (vertical scale) based on a CPU threshold (~70%), and the recorded demo successfully provisions and tears down nodes.

AI-written summary · about

" chapters (5)

1. 0:00Start
2. 13:00The Nomad Autoscaler
3. 16:51Prometheus
4. 17:59Min and Max
5. 22:43Prometheus for Logs

" related talks

• 01h 06mSchedule AI workloads efficiently on Nomad with node poolsTu Nguyen · HashiConf 2025
• 27mNomad Best Practices for Reliable DeploysBram Vogelaar, Srinivas Vadlamani · HashiTalks 2020
• 22mAdding Observability to HashiCorp Nomad Applications with GrafanaCyril Tovena · HashiTalks 2021
• 44mNomad, orchestration made in HashiCorpBastien Cadiot · Paris Container Day 2017
• 35mLiving the Nomadic lifeNic Jackson · Paris Container Day 2017
• 46mProgressive Deployments with Consul Service MeshNic Jackson · HashiConf Europe 2022
• 21mMonitoring Nomad with Prometheus and IcingaBram Vogelaar · HashiTalks 2020
• 30mA beginner's journey: Scaling self-hosted platform infrastructure with NomadBenjamin Lykins · HashiConf 2024