HPA based on CPU and Memory utilization

Mon, 16 Mar 2020 00:00:00 +0000

In this post, I will talk about the HPA, which stands for “Horizontal Pod Autoscaler”. It is a Kubernetes resource that automatically adjusts the number of replicas (pods) in a deployment based on observed CPU or memory utilization.

Why use an HPA

The goal of an HPA is to make sure your application has enough resources to handle demand. If resource utilization goes up, the HPA increases the number of replicas (more pods running) so the load is properly distributed. Likewise, if utilization goes down, the HPA reduces the number of replicas, saving resources. This keeps your application scalable and performant automatically and efficiently.

Make sure to replace the name “meu-hpa” with the name you want for the HPA, and meu-deployment with the name of the deployment the HPA will be applied to. This ensures the HPA automatically adjusts the number of replicas of your deployment based on memory and CPU usage, keeping them at around 80% utilization.

About

Thu, 28 Feb 2019 00:00:00 +0000

~/luks.dev.br

$ whoami
lucas-oliveira-dantas
$ cat profile.yaml
role:       DevOps · SRE · Cloud
focus:      DevSecOps and software supply chain security
experience: 10+ years in IT
based in:   Rio Grande do Norte, Brazil
blog:       writing since 2019
lattes:     lattes.cnpq.br/8767989273085683

Hi! 👋 I’m Lucas Oliveira, a technology professional with more than a decade of experience in DevOps, SRE and Cloud. My job is getting software to production fast, safely and reliably: pipelines, containers, Kubernetes, infrastructure as code and, more and more, security across all of those stages.

I started like many people in this field: spending hours in front of the computer, creating virtual machines and hosting simple pages just for the joy of seeing them work. That enthusiasm led me to AWS, Azure and GCP, to the cloud native ecosystem and, more recently, to DevSecOps, where most of my studies are focused today.

This blog is where I document what I learn in practice: real problems, tested solutions and the reasoning behind them. Most articles are written in Portuguese; you can read them through the machine translation links at the top of each post.

What I do

🛡️ DevSecOps

SAST with Semgrep and SonarQube, image scanning with Docker Scout, SBOM, SLSA and secrets management with Vault and Terraform.

🔁 CI/CD & GitOps

Pipelines on GitLab CI, GitHub Actions, Azure DevOps and Harness, with reusable templates and well-designed cache and artifacts.

☸️ Containers & Kubernetes

Docker from Dockerfile to production, autoscaling with HPA, pod priority and cluster troubleshooting.

☁️ Cloud & IaC

AWS (including serverless with SAM), GCP and Azure, with infrastructure versioned in Terraform and Ansible.

Everyday stack

Languages: PythonJavaScriptNode.jsShell
Containers: DockerDocker ComposeKubernetes
CI/CD: GitLab CIGitHub ActionsAzure DevOpsHarness
Cloud & IaC: AWSGCPAzureTerraformAnsible
Security: SemgrepSonarQubeDocker ScoutVault
Observability: PrometheusGrafanaElasticsearch
Data: PostgreSQLMongoDB

Where to start

A few articles that represent what I write about (in Portuguese):

In English: HPA based on CPU and memory utilization.

Academic background

Beyond hands-on practice, I value a solid academic foundation and formal learning. My education and career are recorded in my Lattes CV, the official platform of CNPq (Brazil’s National Council for Scientific and Technological Development) that gathers the academic and professional history of researchers and professionals in Brazil. It’s in Portuguese.

Lattes CV · CNPq lattes.cnpq.br/8767989273085683 ↗

Projects and community

AI-assisted software supply chain security: I’ve been building a tool for vulnerability scanning, SBOM generation and SLSA verification with AI-generated explanations. See a preview on X.
Cloud native community in RN: I follow and take part in the events of Cloud Native RN, the CNCF community group in Rio Grande do Norte.
Open source: examples and scripts from the articles live on GitHub and GitLab.

Let’s talk

Want to chat about DevOps, security or one of the articles here? Reach out (all channels are on the Links page):

LinkedIn Lattes GitHub GitLab X Email

Links

Mon, 01 Jan 0001 00:00:00 +0000

Everywhere I publish, share code or can be reached. To learn more about me, see the About page.

Search