Luiz Carneiro

Build Multi-Agent Systems with ADK Co-organized with GDG Cloud Munich, this workshop brought together an amazing community to explore multi-agent systems 🙌 Participants learned to run, test & deploy agents on Vertex AI. 👏 Thanks to everyone who joined! #BuildWithAI #GDG

Give your agents some tools and let them talk to each other! 🤖 I am #tooling up to hit the stage at @googlecloud Builders Day Berlin next Tuesday, March 17 to talk exactly about that: When #A2A 🤖🤝🤖 meets #MCP 🧰 🎟️: cloud.google.com/events/builder… See you in Berlin! 🐻

Just continue. #Kubernetes #31

Not directly to Kubernetes related, but to #trustability, I found out yesterday that the old “Turbo” button in accent machines used to make the computers slower instead of turbo 🤣 #reliability #kubernetes #30

@GooglePrishtina @gdgcloudmunich Let’s do it 🚀

🚀 One Week Left! Join our online workshop: Build Multi-Agent Systems with ADK! Led by Luiz Carneiro, this workshop is beginner-friendly and practical. If you’re curious about AI agents and building real systems → don’t miss it. #BuilWithAI

@OneUptimeHQ Rocked it! 🪨

@carneiroDotDev The missing piece most teams hit: connecting these three signals during an incident. Having logs, metrics, and traces in separate tools means context-switching when you can least afford it. Correlated observability is where reliability actually gets built.

Logs tell you what happened. Metrics tell you how it's performing. Traces tell you where the delay is. Visibility is the prerequisite for reliability. #28 #Kubernetes ☸️ #K8s #Observability #SRE #CloudNative

The Reconciliation Loop: Current State vs. Desired State. If they don't match, the controller acts. Continuous correction is what makes the system self-healing. #29 #kubernetes ☸️ #K8s #DevOps #CloudNative #SRE

Kubelet receives PodSpecs and ensures the described containers are running and healthy. It acts as the bridge between the Control Plane's intent and the Node's reality. Without Kubelet, the cluster cannot execute. #27 #kubernetes ☸️ #K8s #SRE #CloudNative

The Control Plane: The brain of #Kubernetes. 🧠 It manages the cluster state, makes scheduling decisions, and responds to events. #26 #kubernetes #K8s #DevOps #CloudNative

Gateway API: Transitioning from Ingress to intent-based networking. 🛣️ #25 #kubernetes #K8s #Networking #GatewayAPI #DevOps

etcd: The source of truth for the cluster. 📀 This distributed key-value store maintains every object's state. If etcd fails, the cluster loses its memory. #24 #kubernetes ☸️ #K8s #SRE #Architecture #CloudNative

ReferenceGrant enables a Gateway to safely reference Secrets or Services in different namespaces. #23 ☸️ #K8s #GatewayAPI #Security #kubernetes

Admission Controllers intercept requests to the API server to validate or mutate objects before persistence. Enforce security policies and governance at the source. #22☸️ #kubernetes #K8s #Security #DevOps #CloudNative

⚠️ Warning: Ingress NGINX API deprecation begins in March. Migrate to the Gateway API to decouple routing from infrastructure.☸️ #K8s #Networking #GatewayAPI #DevOps

Service Mesh: The dedicated layer for service-to-service communication. 🕸️ It offloads networking logic from the application code to the infrastructure. #20 #Kubernetes☸️ #K8s #ServiceMesh #CloudNative #DevOps

Wise words from lil man.

Network Policies: The internal firewalls of #Kubernetes. 🛡️ Default-allow is a risk. Zero-trust architecture requires explicit egress and ingress rules to isolate traffic at the Pod level. #19 #Kubernetes ☸️ #K8s #Security #DevOps #CloudNative

Pod Disruption Budgets (PDB): The uptime insurance. 🛡️ PDBs ensure a minimum number of replicas remain available during voluntary disruptions, like node upgrades. #18 #Kubernetes ☸️ #K8s #DevOps #SRE #CloudNative