Autoscaling Mesos Cluster
summary
A small Python autoscaler is built that polls Mesos leader / slaves endpoint every configurable interval, computes cluster CPU usage and calls a provider class to grow or shrink Azure VM Scale Sets; default trigger is 80% CPU, min 3 and max 20 nodes. The CLI uses Click, the code is ~200 lines, packaged in Docker, and the demo scales a stateless agent pool from 3 to 6 instances successfully. Source is published and provider abstraction lets it target other clouds; stateful workloads are excluded.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:15Agenda
- 1:47Context
- 3:27Marketing
- 5:27SES Engine
- 7:38VM Skill Set
- 10:35Code
- 11:26Demo
- 12:30Options
- 14:30Deploy
- 17:06Autoscaling
- 23:05Bootstrapping
- 28:47Tags
related talks
16mAutoscaling Distributed System with BOSH
29mCustom App Autoscaler Using Cloud Controller API - Christopher Decelles
07mAutoscaling with virtual node and Azure Kubernetes Service (AKS)
20mDrive higher utilization of apps in Azure Spring Cloud with Autoscale
57mScaling Azure Container Apps: Maximize Performance and Minimize Costs
01h 04mScaling Docker Containers using Kubernetes and Azure Container Service
20mManage Scaling, While Keeping Costs Under Control. Challenge Accepted.
17mAutomatic compute scaling with serverless for Hyperscale in Azure SQL Databases