Vertical and horizontal autoscaling on Kubernetes Engine

preview_player
Показать описание

Critter Junction’s traffic is set to increase exponentially, and their current infrastructure can’t meet their traffic needs. In this episode of Season of Scale, Carter Morgan shows you how to leverage Google Kubernetes Engine vertical and horizontal autoscaling to better manage workloads, or to modify resources provisioned to an individual service. Watch to learn how vertical and horizontal scaling can help you make easily scalable applications!

0:00 - Intro
0:47 - Horizontal Pod Autoscaling
1:43 - Cluster Autoscaling
2:10 - Horizontal vs. Vertical scaling
2:37 - Vertical Pod Autoscaling
3:18 - Conclusion

#SeasonofScale

Product: Application Architecture, HPA, VPA, CA; fullname: Carter Morgan;
Рекомендации по теме
Комментарии
Автор

When we are using VerticalPodScaling to adjust POD's CPU/memory utilization during run time, then why was the UpdateMode set to off?

nikhilgupta
Автор

Wow, great!!! However a question: what are the trigger to vertical scale up and down the pods? are the same as for the horizontal scale (memory, cpu, custom)? If so, what is the strategy? Does the pod start to scale up, and then (when the scale up limits are reached) scale out?

guillaumeblaquiere