Description

Next level leveraging cloud computing

We moved 70% of our services to the cloud and new services start in the cloud unless... Time to dive into a topic brought up by our tech leads. Want to know how you can maximize your cloud benefits?

Alexandra wrote a blog post on Horizontal Pod Auto-scaling, HPA.

What this episode covers

It triggered us for this episode.

Maximizing cloud benefits can be done in many different ways. For this episode, we dive into auto-scaling. Cloud computing brings great features for scaling up and down based on the needs of an application. But if you think my application can scale on-premise, just copy that principle in the cloud. Think again. You might be overspending or still underperforming. Our guests explain why this is and what to do to maximize your cloud benefits.

Topics discussed

• Introduction
• What our cloud strategy is (GCP, run in Kubenetes, Istio, cloud-native, etc)
• Migration strategies for feature teams (check)
  • Lift and shift
  • Move and improve
  • Rip and replace
• Explain horizontal vs vertical scaling (based on the lawnmower metaphor)
• A concrete example of using k8s HPA: MESS application
  • Prerequisites before setting up HPA
  • Before and after setup:
    • The only way to make this faster is by adding bigger hardware and remain the number of instances.
    • And IF you can add more instances easily, then there is the problem that the new instance contains everything, so you’re not only scaling the parts where it hurts the most.
• Cost optimization

• Rogier Lommers – Space Tech Lead at bol.com
• Alexandra Egher - Product Tech Lead at bol.com

and both friends of the show!

• The podcast about Axle, our own framework for developers based on Spring Boot can be found here 
• R2D2, our own cloud integration tool is explained in this episode
• Link to Techlab blog post of Alexandra

Sidenote

In general, a story about architecture that relates to all of this.