
clovis
3.6K posts

clovis
@clovisdsdo
Chasing Kubernetes wisdom • DevOps Engineer • laC addict • Building platforms
Houston, TX Katılım Mart 2011
1.4K Takip Edilen2.8K Takipçiler
Sabitlenmiş Tweet

I asked my AI agent to audit my Kubernetes namespace security using this prompt:
Analyze the security posture of my namespace. Identify risks, explain them, and suggest fixes.
It found:
- No NetworkPolicies
- Public API server open to 0.0.0.0/0
- Pods using nginx:latest
- Default service account
- No IAM roles (IRSA)
And gave me the exact fixes.
Watch the video 👇👇👇👇
English

Kubernetes Tools Ecosystem ⚙️
A visual map of the most popular tools used around Kubernetes.
→ Cluster Management
EKS • AKS • GKE • Rancher • Helm • Minikube
→ Networking
Calico • Cilium • Flannel • Istio • Traefik
→ Infra Automation
Terraform • Ansible • Pulumi • ArgoCD • Flux
→ Container Runtime
Docker • containerd • CRI-O • gVisor • Kata
→ Security
Falco • Vault • Trivy • Kyverno • OPA
→ Monitoring & Observability
Prometheus • Grafana • ELK • Jaeger • Datadog
Kubernetes is powerful on its own.
But the real ecosystem around it is what makes it production-ready.

Deutsch

@Pranj_Sometimes Good job 👏
I assume this was done sequentially: 1.33 -> 1.34 —> 1.35.
English

Upgraded 22+ Amazon EKS clusters from v1.33 to v1.35 with zero downtime 🚀
And yes, the complete upgrade was handled using Terraform, no manual changes.
Before starting any EKS upgrade, I always spend time validating the cluster properly instead of directly upgrading.
Things I checked first:
• Upgrade Insights
• Version skew issues
• Deprecated APIs
• Existing cluster warnings/errors
• Workload health and node readiness
If there are any issues, fix them first before proceeding.
Next step was compatibility validation through AWS official documentation for:
• CoreDNS
• kube-proxy
• VPC CNI
• EBS CSI Driver
• Metrics Server
• AWS Load Balancer Controller
Since the cluster was using Karpenter, I upgraded Karpenter first to a version compatible with EKS v1.35 before touching the cluster upgrade.
Upgrade approach:
✅ Validated upgrade insights
✅ Verified addon compatibility
✅ Upgraded Karpenter version
✅ Used Terraform for EKS and addon upgrades
✅ Gradually upgraded managed node groups
✅ Monitored workloads continuously using k9s
✅ Validated pods, ingress, monitoring, and applications after every step
Biggest learning:
Kubernetes upgrades are not difficult when planning, compatibility checks, automation, and monitoring are done correctly.
#Terraform #Kubernetes #EKS #AWS #DevOps #Karpenter #IaC #PlatformEngineering #SRE #k9s

English

Behind YouTube, there is a story most people don’t know. Before videos became easy to upload, search, and share, one Super Bowl incident showed the world that the internet was missing something important.
1️⃣February 1, 2004: During the Super Bowl XXXVIII halftime show, Justin Timberlake exposed part of Janet Jackson’s breast on live TV. The moment became one of the most searched and replayed media incidents of that time.
2️⃣After February 2004: Many people wanted to find the clip online, but video sharing on the internet was still difficult. There was no simple platform like YouTube yet.
3️⃣2004: YouTube co-founder Jawed Karim later said the Janet Jackson incident was one of the moments that showed the need for an easy video-sharing website. Another example often mentioned was difficulty finding videos from the 2004 Indian Ocean tsunami.
4️⃣February 14, 2005: The domain YouTube.comwas activated/registered. The founders were Chad Hurley, Steve Chen, and Jawed Karim.
5️⃣April 23, 2005: The first YouTube video, “Me at the zoo,” was uploaded by Jawed Karim. This is considered the first video in YouTube history.
6️⃣April/May 2005: YouTube opened to users in beta form. The original idea was not fully clear at first, but the platform quickly became focused on uploading and sharing videos easily.
7️⃣December 15, 2005: YouTube officially launched out of beta. By then, it was already growing fast.
8️⃣October/November 2006: Google acquired YouTube for $1.65 billion in stock, turning it into one of the biggest internet acquisitions of that era.

English

@clovisdsdo I do believe companies are just laying off for the sake of it now, everyone doing it so we also need to do it
English

@clovisdsdo First: breathe. Then check if your HPA is fighting the Cluster Autoscaler — seen that deadlock more times than I can count.
English

Kubernetes Incident Response
It is 9:00 AM on a Friday
Your company just launched a major promotion, and traffic suddenly spikes by 500%
Users start reporting that checkout is timing out
At the same time, monitoring shows a spike in HTTP 504 Gateway Timeout errors
You check the Kubernetes cluster and notice that the order-processor Pods are crashing one by one
The HPA is trying to create more Pods
The Cluster Autoscaler is trying to add new worker nodes
But the cluster feels completely locked up
As the DevOps/SRE engineer on call, walk me through how you would t fix this live production incident
English


@cloclodma $0 for a project that now runs the entire tech industry is crazy.
English





