I Fixed Kubernetes Autoscaling using Machine Learning | ft. Keda & Prophet
Explore how you can predictively scale your workloads to reduce downtime.
The primary way of autoscaling microservices in Kubernetes is by using an HPA. Having said that, for most modern workloads, the Horizontal Pod Autoscaler isn't enough. We need to predictively scale our workloads based on historic patterns to be ready for peak load periods.
In this video we will:
1] Decide how to select input and output parameters for our machine learning model.
2] Train a model using Facebook Prophet.
3] Integrate the model with Kubernetes using Keda
4] Discuss some helpful tips for implementing predictive autoscaling in production.
--------------------------------
References:
1] The Code: https://github.com/YourTechBud/ytb-pr...
2] Keda Website: https://keda.sh/
3] Facebook Prophet: https://facebook.github.io/prophet/
--------------------------------
Related Videos:
1] Why You Shouldn't Use K8s Autoscaling? - • Why You Shouldn't Use K8s Autoscaling?!!
2] Why your Microservices needs Kubernetes?! - • Why your Microservices needs Kubernet...
--------------------------------
Chapters:
00:00 - Introduction
01:22 - Deciding the Input & Output
03:35 - Training the model
06:06 - Figuring out Inference
07:04 - Integrating with Kubernetes
08:23 - Tips for Production
#Keda #Kubernetes #Autoscaling #MachineLearning #DevOps #Microservices
Watch video I Fixed Kubernetes Autoscaling using Machine Learning | ft. Keda & Prophet online without registration, duration hours minute second in high quality. This video was added by user YourTechBud Codes 14 November 2023, don't forget to share it with your friends and acquaintances, it has been viewed on our site 1,501 once and liked it 39 people.