Building an Agentic AI Platform on EKS

Versiones compatibles: EKS 1.31+, vLLM 0.6+, Karpenter 1.0+ Última actualización: February 23, 2026

La Agentic AI va más allá de las respuestas simples a preguntas: crea planes de forma autónoma, usa herramientas y alcanza objetivos de manera iterativa. Este capítulo cubre cómo construir una plataforma de Agentic AI de nivel de producción en EKS.

1. Agentic AI Platform Overview

What is Agentic AI?

La Agentic AI es un sistema de IA autónomo con las siguientes características:

1. Planificación autónoma: Descompone tareas complejas en subtareas y determina el orden de ejecución.
2. Ejecución basada en herramientas: Utiliza varias herramientas, incluidas API externas, bases de datos y ejecutores de código.
3. Mejora iterativa: Evalúa los resultados de ejecución y modifica los planes según sea necesario.
4. Gestión de estado: Mantiene estado y memoria para tareas de larga duración.

Why Kubernetes is Needed

Kubernetes proporciona las siguientes capacidades principales para plataformas de Agentic AI:

Requirement	Kubernetes Solution
GPU Orchestration	Device Plugin, GPU Operator, MIG
Auto Scaling	HPA, VPA, Karpenter
Multi-tenant Isolation	Namespace, NetworkPolicy, ResourceQuota
High Availability	ReplicaSet, PodDisruptionBudget
Service Mesh	Istio, Gateway API
Cost Optimization	Spot instances, Node consolidation

Four Key Technical Challenges

Desafíos clave que se deben resolver al construir una plataforma de Agentic AI:

---

2. GPU Infrastructure Configuration

GPU Instance Type Comparison

Principales tipos de instancias GPU disponibles en AWS:

Instance	GPU	GPU Memory	Use Case	Hourly Cost (On-Demand)
p5.48xlarge	8x H100	640GB	Large-scale training, very large model inference	~$98.32
p4d.24xlarge	8x A100	320GB	Distributed training, 70B+ model inference	~$32.77
g5.xlarge	1x A10G	24GB	Small-medium model inference	~$1.01
g5.48xlarge	8x A10G	192GB	Multi-model serving	~$16.29
g6.xlarge	1x L4	24GB	Cost-effective inference	~$0.80
g6.48xlarge	8x L4	192GB	Large-scale inference cluster	~$13.35
inf2.xlarge	1x Inferentia2	32GB	AWS-optimized inference	~$0.76

Multi-Instance GPU (MIG) Configuration

Las GPU NVIDIA A100/H100 se pueden particionar físicamente mediante MIG para aislar múltiples cargas de trabajo.

MIG Profiles (A100 80GB Reference)

Profile	GPU Memory	SM Count	Use Case
1g.10gb	10GB	14	Small model inference, development
2g.20gb	20GB	28	7B model inference
3g.40gb	40GB	42	13B model inference
4g.40gb	40GB	56	Large batch inference
7g.80gb	80GB	98	70B model, training

NVIDIA GPU Operator Deployment

# gpu-operator-values.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: gpu-operator-config
  namespace: gpu-operator
data:
  mig.strategy: "mixed"  # single or mixed
---
apiVersion: helm.toolkit.fluxcd.io/v2
kind: HelmRelease
metadata:
  name: gpu-operator
  namespace: gpu-operator
spec:
  interval: 10m
  chart:
    spec:
      chart: gpu-operator
      version: "v24.9.0"
      sourceRef:
        kind: HelmRepository
        name: nvidia
        namespace: flux-system
  values:
    operator:
      defaultRuntime: containerd
    mig:
      strategy: mixed
    devicePlugin:
      enabled: true
      config:
        name: time-slicing-config
        default: any
    gfd:
      enabled: true
    dcgmExporter:
      enabled: true
      serviceMonitor:
        enabled: true

# GPU Operator installation
helm repo add nvidia https://helm.ngc.nvidia.com/nvidia
helm repo update

helm install gpu-operator nvidia/gpu-operator \
  --namespace gpu-operator \
  --create-namespace \
  --values gpu-operator-values.yaml

# Check MIG configuration
kubectl get nodes -l nvidia.com/mig.capable=true \
  -o jsonpath='{range .items[*]}{.metadata.name}: {.status.allocatable}{"\n"}{end}'

MIG Partition Configuration

# mig-config.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: mig-parted-config
  namespace: gpu-operator
data:
  config.yaml: |
    version: v1
    mig-configs:
      # Development environment: Small partitions for many users
      development:
        - devices: [0]
          mig-enabled: true
          mig-devices:
            "1g.10gb": 7

      # Production: Medium-sized partitions
      production-inference:
        - devices: [0]
          mig-enabled: true
          mig-devices:
            "2g.20gb": 3
            "1g.10gb": 1

      # Large models: Full GPU usage
      large-model:
        - devices: [0]
          mig-enabled: true
          mig-devices:
            "7g.80gb": 1

Time-Slicing Configuration

Para las GPU que no admiten MIG (A10G, L4, etc.), Time-Slicing permite compartir GPU.

# time-slicing-config.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: time-slicing-config
  namespace: gpu-operator
data:
  any: |
    version: v1
    flags:
      migStrategy: none
    sharing:
      timeSlicing:
        renameByDefault: false
        failRequestsGreaterThanOne: false
        resources:
          - name: nvidia.com/gpu
            replicas: 4  # Split one GPU into 4
---
# Apply Time-Slicing to node
apiVersion: v1
kind: Node
metadata:
  name: gpu-node-1
  labels:
    nvidia.com/device-plugin.config: time-slicing-config

MIG vs Time-Slicing Comparison

Characteristic	MIG	Time-Slicing
Isolation Level	Hardware isolation (memory, SM)	Software isolation (time division)
Supported GPUs	A100, H100	All NVIDIA GPUs
Memory Guarantee	Guaranteed	Shared (contention possible)
Overhead	Low	Context switching overhead
Flexibility	Requires reconfiguration	Dynamically adjustable
Use Cases	Production, multi-tenant	Development, batch processing

Karpenter NodePool Configuration

# gpu-nodepool.yaml
apiVersion: karpenter.sh/v1
kind: NodePool
metadata:
  name: gpu-inference
spec:
  template:
    metadata:
      labels:
        workload-type: inference
    spec:
      requirements:
        - key: kubernetes.io/arch
          operator: In
          values: ["amd64"]
        - key: karpenter.sh/capacity-type
          operator: In
          values: ["on-demand", "spot"]
        - key: node.kubernetes.io/instance-type
          operator: In
          values:
            - g5.xlarge
            - g5.2xlarge
            - g5.4xlarge
            - g6.xlarge
            - g6.2xlarge
        - key: "karpenter.k8s.aws/instance-gpu-count"
          operator: Gt
          values: ["0"]
      nodeClassRef:
        group: karpenter.k8s.aws
        kind: EC2NodeClass
        name: gpu-nodes
      taints:
        - key: nvidia.com/gpu
          value: "true"
          effect: NoSchedule
  limits:
    nvidia.com/gpu: 100
  disruption:
    consolidationPolicy: WhenEmptyOrUnderutilized
    consolidateAfter: 5m
    budgets:
      - nodes: "20%"
---
apiVersion: karpenter.k8s.aws/v1
kind: EC2NodeClass
metadata:
  name: gpu-nodes
spec:
  amiFamily: AL2023
  role: KarpenterNodeRole-${CLUSTER_NAME}
  subnetSelectorTerms:
    - tags:
        karpenter.sh/discovery: ${CLUSTER_NAME}
  securityGroupSelectorTerms:
    - tags:
        karpenter.sh/discovery: ${CLUSTER_NAME}
  blockDeviceMappings:
    - deviceName: /dev/xvda
      ebs:
        volumeSize: 200Gi
        volumeType: gp3
        iops: 10000
        throughput: 500
        deleteOnTermination: true
  tags:
    Environment: production
    Workload: ai-inference

---

3. Model Serving (vLLM)

vLLM Architecture

vLLM proporciona inferencia LLM de alto rendimiento mediante las siguientes tecnologías principales:

vLLM Deployment Configuration

# vllm-deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: vllm-llama3-70b
  namespace: ai-inference
  labels:
    app: vllm
    model: llama3-70b
spec:
  replicas: 2
  selector:
    matchLabels:
      app: vllm
      model: llama3-70b
  template:
    metadata:
      labels:
        app: vllm
        model: llama3-70b
    spec:
      nodeSelector:
        workload-type: inference
      tolerations:
        - key: nvidia.com/gpu
          operator: Equal
          value: "true"
          effect: NoSchedule
      containers:
        - name: vllm
          image: vllm/vllm-openai:v0.6.4
          ports:
            - containerPort: 8000
              name: http
          env:
            - name: HUGGING_FACE_HUB_TOKEN
              valueFrom:
                secretKeyRef:
                  name: huggingface-token
                  key: token
            - name: VLLM_ATTENTION_BACKEND
              value: "FLASH_ATTN"
          args:
            - "--model"
            - "meta-llama/Meta-Llama-3-70B-Instruct"
            - "--tensor-parallel-size"
            - "4"
            - "--gpu-memory-utilization"
            - "0.95"
            - "--max-model-len"
            - "8192"
            - "--enable-prefix-caching"
            - "--enable-chunked-prefill"
            - "--max-num-batched-tokens"
            - "32768"
            - "--trust-remote-code"
          resources:
            requests:
              nvidia.com/gpu: 4
              memory: "200Gi"
              cpu: "32"
            limits:
              nvidia.com/gpu: 4
              memory: "250Gi"
              cpu: "48"
          readinessProbe:
            httpGet:
              path: /health
              port: 8000
            initialDelaySeconds: 300
            periodSeconds: 10
            timeoutSeconds: 5
            failureThreshold: 3
          livenessProbe:
            httpGet:
              path: /health
              port: 8000
            initialDelaySeconds: 600
            periodSeconds: 30
            timeoutSeconds: 10
            failureThreshold: 3
          volumeMounts:
            - name: model-cache
              mountPath: /root/.cache/huggingface
            - name: shm
              mountPath: /dev/shm
      volumes:
        - name: model-cache
          persistentVolumeClaim:
            claimName: model-cache-pvc
        - name: shm
          emptyDir:
            medium: Memory
            sizeLimit: 64Gi
---
apiVersion: v1
kind: Service
metadata:
  name: vllm-llama3-70b
  namespace: ai-inference
spec:
  selector:
    app: vllm
    model: llama3-70b
  ports:
    - port: 8000
      targetPort: 8000
      name: http
  type: ClusterIP

Performance Optimization Settings

Tensor Parallelism

Distribución de modelos grandes entre múltiples GPU:

# Recommended settings by model size
# 7B model: 1 GPU
# 13B model: 1-2 GPU
# 70B model: 4 GPU (A100) or 8 GPU (A10G)
# 405B model: 8 GPU (H100)

args:
  - "--tensor-parallel-size"
  - "4"  # Adjust to match GPU count

KV Cache Management

args:
  # Allocate 95% of GPU memory to KV cache
  - "--gpu-memory-utilization"
  - "0.95"

  # Block size setting (default: 16)
  - "--block-size"
  - "16"

  # Swap space setting (CPU memory)
  - "--swap-space"
  - "32"  # In GB

Prefix Caching

Cache para prompts de sistema repetidos:

args:
  - "--enable-prefix-caching"

# Effect: Reduces Time To First Token (TTFT) by 50-80%
# for requests using identical system prompts

Chunked Prefill

Optimización del procesamiento de contextos largos:

args:
  - "--enable-chunked-prefill"
  - "--max-num-batched-tokens"
  - "32768"

# Effect: Stabilizes response latency for workloads
# with mixed long and short prompts

Model Serving Patterns

Single Model Pod

# Simplest pattern: One Pod serving one model
spec:
  containers:
    - name: vllm
      args:
        - "--model"
        - "meta-llama/Meta-Llama-3-8B-Instruct"

Disaggregated Serving with llm-d

Separación de Prefill y Decode para optimización:

# llm-d-prefill.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: llm-d-prefill
spec:
  replicas: 2
  template:
    spec:
      containers:
        - name: llm-d
          image: llm-d/prefill:latest
          args:
            - "--role"
            - "prefill"
            - "--model"
            - "meta-llama/Meta-Llama-3-70B-Instruct"
          resources:
            requests:
              nvidia.com/gpu: 4
---
# llm-d-decode.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: llm-d-decode
spec:
  replicas: 4
  template:
    spec:
      containers:
        - name: llm-d
          image: llm-d/decode:latest
          args:
            - "--role"
            - "decode"
            - "--prefill-endpoint"
            - "http://llm-d-prefill:8000"
          resources:
            requests:
              nvidia.com/gpu: 2

---

4. Inference Gateway

Gateway API-based AI Workload Routing

Extensión de Kubernetes Gateway API para enrutar de forma eficiente cargas de trabajo de inferencia de IA.

Kgateway + InferencePool Architecture

InferencePool CRD

# inferencepool.yaml
apiVersion: inference.networking.x-k8s.io/v1alpha1
kind: InferencePool
metadata:
  name: llama3-pool
  namespace: ai-inference
spec:
  targetPortNumber: 8000
  selector:
    matchLabels:
      app: vllm
      model: llama3-70b
  endpointPickerConfig:
    # Load balancing strategy
    extensionRef:
      name: prefix-aware-picker
      group: inference.networking.x-k8s.io
      kind: EndpointPicker
---
apiVersion: inference.networking.x-k8s.io/v1alpha1
kind: EndpointPicker
metadata:
  name: prefix-aware-picker
  namespace: ai-inference
spec:
  type: PrefixAware
  config:
    # Prefix cache hit rate optimization
    prefixHashBuckets: 1024
    fallbackStrategy: LeastLoaded
    loadMetric: pending_requests
---
apiVersion: gateway.networking.k8s.io/v1
kind: HTTPRoute
metadata:
  name: llama3-route
  namespace: ai-inference
spec:
  parentRefs:
    - name: ai-gateway
      namespace: ai-inference
  rules:
    - matches:
        - path:
            type: PathPrefix
            value: /v1/chat/completions
          headers:
            - name: x-model
              value: llama3-70b
      backendRefs:
        - group: inference.networking.x-k8s.io
          kind: InferencePool
          name: llama3-pool
          port: 8000

LiteLLM Integrated Gateway

LiteLLM unifica varios proveedores de LLM bajo una única API.

# litellm-deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: litellm-gateway
  namespace: ai-gateway
spec:
  replicas: 3
  selector:
    matchLabels:
      app: litellm
  template:
    metadata:
      labels:
        app: litellm
    spec:
      containers:
        - name: litellm
          image: ghcr.io/berriai/litellm:main-v1.55.0
          ports:
            - containerPort: 4000
          env:
            - name: LITELLM_MASTER_KEY
              valueFrom:
                secretKeyRef:
                  name: litellm-secrets
                  key: master-key
            - name: DATABASE_URL
              valueFrom:
                secretKeyRef:
                  name: litellm-secrets
                  key: database-url
          volumeMounts:
            - name: config
              mountPath: /app/config.yaml
              subPath: config.yaml
          resources:
            requests:
              cpu: "2"
              memory: "4Gi"
            limits:
              cpu: "4"
              memory: "8Gi"
      volumes:
        - name: config
          configMap:
            name: litellm-config
---
apiVersion: v1
kind: ConfigMap
metadata:
  name: litellm-config
  namespace: ai-gateway
data:
  config.yaml: |
    model_list:
      # Internal vLLM endpoints
      - model_name: llama3-70b
        litellm_params:
          model: openai/meta-llama/Meta-Llama-3-70B-Instruct
          api_base: http://vllm-llama3-70b.ai-inference:8000/v1
          api_key: dummy
        model_info:
          max_tokens: 8192
          input_cost_per_token: 0.0000001
          output_cost_per_token: 0.0000003

      - model_name: llama3-8b
        litellm_params:
          model: openai/meta-llama/Meta-Llama-3-8B-Instruct
          api_base: http://vllm-llama3-8b.ai-inference:8000/v1
          api_key: dummy
        model_info:
          max_tokens: 8192
          input_cost_per_token: 0.00000005
          output_cost_per_token: 0.00000015

      # External providers (for fallback)
      - model_name: gpt-4o
        litellm_params:
          model: gpt-4o
          api_key: os.environ/OPENAI_API_KEY
        model_info:
          max_tokens: 128000
          input_cost_per_token: 0.000005
          output_cost_per_token: 0.000015

      - model_name: claude-3-5-sonnet
        litellm_params:
          model: anthropic/claude-3-5-sonnet-20241022
          api_key: os.environ/ANTHROPIC_API_KEY
        model_info:
          max_tokens: 200000
          input_cost_per_token: 0.000003
          output_cost_per_token: 0.000015

    # Routing settings
    router_settings:
      routing_strategy: usage-based-routing-v2
      enable_pre_call_checks: true
      redis_host: redis.ai-gateway
      redis_port: 6379

    # Fallback settings
    litellm_settings:
      fallbacks:
        - model: llama3-70b
          fallback_models:
            - gpt-4o
            - claude-3-5-sonnet

      # Retry settings
      num_retries: 3
      request_timeout: 300

      # Cost tracking
      success_callback: ["langfuse"]
      failure_callback: ["langfuse"]
---
apiVersion: v1
kind: Service
metadata:
  name: litellm-gateway
  namespace: ai-gateway
spec:
  selector:
    app: litellm
  ports:
    - port: 4000
      targetPort: 4000
  type: ClusterIP

LiteLLM Usage Example

# litellm_client.py
from openai import OpenAI

# Using LiteLLM gateway
client = OpenAI(
    api_key="sk-litellm-master-key",
    base_url="http://litellm-gateway.ai-gateway:4000/v1"
)

# Call internal model (automatic routing)
response = client.chat.completions.create(
    model="llama3-70b",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Explain Kubernetes in simple terms."}
    ],
    max_tokens=500
)

print(response.choices[0].message.content)

# Automatic fallback to external providers if needed
# (llama3-70b failure -> gpt-4o -> claude-3-5-sonnet in order)

---

5. RAG Data Layer

Milvus Vector Database

Milvus es una base de datos open-source para búsquedas vectoriales a gran escala.

Milvus Operator Deployment

# Install Milvus Operator
helm repo add milvus https://zilliztech.github.io/milvus-helm
helm repo update

helm install milvus-operator milvus/milvus-operator \
  --namespace milvus-system \
  --create-namespace

# milvus-cluster.yaml
apiVersion: milvus.io/v1beta1
kind: Milvus
metadata:
  name: milvus-cluster
  namespace: ai-data
spec:
  mode: cluster
  dependencies:
    etcd:
      inCluster:
        values:
          replicaCount: 3
          persistence:
            enabled: true
            size: 50Gi
    pulsar:
      inCluster:
        values:
          components:
            autorecovery: false
          proxy:
            replicaCount: 2
          broker:
            replicaCount: 2
    storage:
      inCluster:
        values:
          mode: distributed
          fullnameOverride: milvus-minio
          persistence:
            enabled: true
            size: 500Gi
  components:
    # Query Node - Vector search processing
    queryNode:
      replicas: 3
      resources:
        requests:
          cpu: "4"
          memory: "16Gi"
        limits:
          cpu: "8"
          memory: "32Gi"

    # Index Node - Index building (GPU accelerated)
    indexNode:
      replicas: 2
      resources:
        requests:
          cpu: "4"
          memory: "16Gi"
          nvidia.com/gpu: 1
        limits:
          cpu: "8"
          memory: "32Gi"
          nvidia.com/gpu: 1

    # Data Node - Data processing
    dataNode:
      replicas: 2
      resources:
        requests:
          cpu: "2"
          memory: "8Gi"
        limits:
          cpu: "4"
          memory: "16Gi"

    # Proxy - API gateway
    proxy:
      replicas: 2
      serviceType: ClusterIP
      resources:
        requests:
          cpu: "2"
          memory: "4Gi"
        limits:
          cpu: "4"
          memory: "8Gi"
  config:
    common:
      gracefulTime: 30000
    queryNode:
      gracefulTime: 30000

Collection Schema Design

# milvus_schema.py
from pymilvus import (
    connections, Collection, FieldSchema,
    CollectionSchema, DataType, utility
)

# Connect to Milvus
connections.connect(
    alias="default",
    host="milvus-cluster-proxy.ai-data",
    port="19530"
)

# Document collection schema
fields = [
    FieldSchema(name="id", dtype=DataType.VARCHAR, max_length=64, is_primary=True),
    FieldSchema(name="content", dtype=DataType.VARCHAR, max_length=65535),
    FieldSchema(name="embedding", dtype=DataType.FLOAT_VECTOR, dim=1536),
    FieldSchema(name="metadata", dtype=DataType.JSON),
    FieldSchema(name="created_at", dtype=DataType.INT64),
]

schema = CollectionSchema(
    fields=fields,
    description="Document embeddings for RAG"
)

# Create collection
collection = Collection(
    name="documents",
    schema=schema,
    using="default"
)

# Create index
index_params = {
    "metric_type": "COSINE",
    "index_type": "HNSW",  # Or GPU_IVF_FLAT for GPU acceleration
    "params": {
        "M": 16,
        "efConstruction": 256
    }
}

collection.create_index(
    field_name="embedding",
    index_params=index_params
)

# Load collection
collection.load()

Index Type Comparison

Index Type	Characteristics	Memory Usage	Search Speed	Use Case
FLAT	Exact search	High	Slow	Small scale, accuracy priority
IVF_FLAT	Cluster-based	Medium	Fast	General use
HNSW	Graph-based	High	Very fast	Large scale, speed priority
GPU_IVF_FLAT	GPU accelerated	Medium	Very fast	Very large scale, GPU usage
SCANN	Quantization-based	Low	Fast	Memory-constrained environments

Document Ingestion Pipeline

# document-ingestion-job.yaml
apiVersion: batch/v1
kind: Job
metadata:
  name: document-ingestion
  namespace: ai-data
spec:
  template:
    spec:
      containers:
        - name: ingestion
          image: ai-platform/document-ingestion:latest
          env:
            - name: S3_BUCKET
              value: "my-documents-bucket"
            - name: MILVUS_HOST
              value: "milvus-cluster-proxy.ai-data"
            - name: EMBEDDING_MODEL
              value: "text-embedding-3-large"
            - name: OPENAI_API_KEY
              valueFrom:
                secretKeyRef:
                  name: openai-credentials
                  key: api-key
          resources:
            requests:
              cpu: "4"
              memory: "16Gi"
            limits:
              cpu: "8"
              memory: "32Gi"
          volumeMounts:
            - name: temp-storage
              mountPath: /tmp/documents
      volumes:
        - name: temp-storage
          emptyDir:
            sizeLimit: 100Gi
      restartPolicy: OnFailure

Chunking Strategy Implementation

# chunking_strategies.py
from langchain.text_splitter import (
    RecursiveCharacterTextSplitter,
    TokenTextSplitter
)
from langchain_experimental.text_splitter import SemanticChunker
from langchain_openai import OpenAIEmbeddings

# 1. Fixed-size chunking
def fixed_chunking(text: str, chunk_size: int = 1000, overlap: int = 200):
    splitter = RecursiveCharacterTextSplitter(
        chunk_size=chunk_size,
        chunk_overlap=overlap,
        separators=["\n\n", "\n", ".", "!", "?", ",", " ", ""]
    )
    return splitter.split_text(text)

# 2. Token-based chunking (LLM context window optimization)
def token_chunking(text: str, chunk_size: int = 512, overlap: int = 50):
    splitter = TokenTextSplitter(
        encoding_name="cl100k_base",  # GPT-4 tokenizer
        chunk_size=chunk_size,
        chunk_overlap=overlap
    )
    return splitter.split_text(text)

# 3. Semantic chunking (context preservation optimization)
def semantic_chunking(text: str):
    embeddings = OpenAIEmbeddings(model="text-embedding-3-small")
    splitter = SemanticChunker(
        embeddings=embeddings,
        breakpoint_threshold_type="percentile",
        breakpoint_threshold_amount=95
    )
    return splitter.split_text(text)

# Recommendation: Choose strategy by document type
CHUNKING_STRATEGIES = {
    "code": {"strategy": "fixed", "chunk_size": 2000, "overlap": 400},
    "documentation": {"strategy": "semantic"},
    "chat_logs": {"strategy": "fixed", "chunk_size": 500, "overlap": 100},
    "default": {"strategy": "token", "chunk_size": 512, "overlap": 50}
}

RAG Workflow

# rag_workflow.py
from langchain_openai import ChatOpenAI, OpenAIEmbeddings
from langchain_milvus import Milvus
from langchain.chains import RetrievalQA
from langchain.prompts import PromptTemplate

# Vector store connection
embeddings = OpenAIEmbeddings(model="text-embedding-3-large")
vectorstore = Milvus(
    embedding_function=embeddings,
    collection_name="documents",
    connection_args={
        "host": "milvus-cluster-proxy.ai-data",
        "port": "19530"
    }
)

# RAG prompt template
RAG_PROMPT = PromptTemplate(
    template="""Answer the question using the following context.
If you cannot find the answer in the context, say "No information available."

Context:
{context}

Question: {question}

Answer:""",
    input_variables=["context", "question"]
)

# LLM setup (using LiteLLM gateway)
llm = ChatOpenAI(
    model="llama3-70b",
    openai_api_base="http://litellm-gateway.ai-gateway:4000/v1",
    openai_api_key="sk-litellm-master-key",
    temperature=0.1
)

# RAG chain configuration
qa_chain = RetrievalQA.from_chain_type(
    llm=llm,
    chain_type="stuff",
    retriever=vectorstore.as_retriever(
        search_type="mmr",
        search_kwargs={"k": 5, "fetch_k": 20}
    ),
    chain_type_kwargs={"prompt": RAG_PROMPT},
    return_source_documents=True
)

# Execute query
result = qa_chain.invoke({"query": "How does Pod scheduling work in Kubernetes?"})
print(result["result"])

---

6. AI Agent Deployment (Kagent)

Kagent Overview

Kagent es una herramienta nativa de Kubernetes para la gestión del ciclo de vida de agentes de IA.

Agent CRD Definition

# agent-crd.yaml
apiVersion: kagent.dev/v1alpha1
kind: Agent
metadata:
  name: research-agent
  namespace: ai-agents
spec:
  # LLM backend settings
  llm:
    provider: litellm
    model: llama3-70b
    endpoint: http://litellm-gateway.ai-gateway:4000/v1
    temperature: 0.7
    maxTokens: 4096

  # Agent system prompt
  systemPrompt: |
    You are a research assistant that helps users find and analyze information.
    You have access to the following tools:
    - web_search: Search the web for information
    - document_search: Search internal documents
    - calculator: Perform calculations

    Always cite your sources and provide accurate information.

  # Tool definitions
  tools:
    - name: web_search
      type: http
      spec:
        url: http://search-api.tools:8080/search
        method: POST
        headers:
          Content-Type: application/json

    - name: document_search
      type: milvus
      spec:
        host: milvus-cluster-proxy.ai-data
        port: 19530
        collection: documents
        topK: 5

    - name: calculator
      type: python
      spec:
        code: |
          def calculate(expression: str) -> str:
              return str(eval(expression))

  # Memory settings
  memory:
    type: redis
    config:
      host: redis.ai-agents
      port: 6379
      ttl: 3600

  # Resource limits
  resources:
    requests:
      cpu: "500m"
      memory: "512Mi"
    limits:
      cpu: "2"
      memory: "2Gi"

  # Scaling settings
  replicas: 2
  autoscaling:
    enabled: true
    minReplicas: 2
    maxReplicas: 10
    targetCPUUtilization: 70

LangGraph Workflow Orchestration

Uso de LangGraph para implementar workflows de IA complejos.

# langgraph_workflow.py
from typing import TypedDict, Annotated, Sequence
from langchain_core.messages import BaseMessage, HumanMessage, AIMessage
from langchain_openai import ChatOpenAI
from langgraph.graph import StateGraph, END
from langgraph.prebuilt import ToolExecutor
import operator

# State definition
class AgentState(TypedDict):
    messages: Annotated[Sequence[BaseMessage], operator.add]
    current_step: str
    iteration: int
    max_iterations: int
    tools_output: dict

# LLM setup
llm = ChatOpenAI(
    model="llama3-70b",
    openai_api_base="http://litellm-gateway.ai-gateway:4000/v1",
    openai_api_key="sk-litellm-master-key"
)

# Node functions
def planner(state: AgentState) -> AgentState:
    """Node that creates task plans"""
    messages = state["messages"]

    planning_prompt = """Based on the user's request, create a step-by-step plan.
    Format your response as a numbered list of steps."""

    response = llm.invoke(messages + [HumanMessage(content=planning_prompt)])

    return {
        "messages": [response],
        "current_step": "execute",
        "iteration": state["iteration"]
    }

def executor(state: AgentState) -> AgentState:
    """Node that executes plans"""
    messages = state["messages"]

    execution_prompt = """Execute the current step of the plan.
    If you need to use a tool, specify the tool and parameters."""

    response = llm.invoke(messages + [HumanMessage(content=execution_prompt)])

    return {
        "messages": [response],
        "current_step": "evaluate",
        "iteration": state["iteration"]
    }

def evaluator(state: AgentState) -> AgentState:
    """Node that evaluates results"""
    messages = state["messages"]

    evaluation_prompt = """Evaluate the execution result.
    Respond with either:
    - COMPLETE: if the task is fully done
    - CONTINUE: if more steps are needed
    - RETRY: if the current step needs to be retried"""

    response = llm.invoke(messages + [HumanMessage(content=evaluation_prompt)])

    return {
        "messages": [response],
        "current_step": "route",
        "iteration": state["iteration"] + 1
    }

def router(state: AgentState) -> str:
    """Router that determines next step"""
    last_message = state["messages"][-1].content.upper()

    if state["iteration"] >= state["max_iterations"]:
        return "end"

    if "COMPLETE" in last_message:
        return "end"
    elif "RETRY" in last_message:
        return "execute"
    else:
        return "plan"

# Graph construction
workflow = StateGraph(AgentState)

# Add nodes
workflow.add_node("plan", planner)
workflow.add_node("execute", executor)
workflow.add_node("evaluate", evaluator)

# Add edges
workflow.set_entry_point("plan")
workflow.add_edge("plan", "execute")
workflow.add_edge("execute", "evaluate")
workflow.add_conditional_edges(
    "evaluate",
    router,
    {
        "plan": "plan",
        "execute": "execute",
        "end": END
    }
)

# Compile graph
app = workflow.compile()

# Execute
initial_state = {
    "messages": [HumanMessage(content="Research the latest trends in Kubernetes security")],
    "current_step": "plan",
    "iteration": 0,
    "max_iterations": 5,
    "tools_output": {}
}

result = app.invoke(initial_state)

Multi-Agent Collaboration Patterns

Supervisor Pattern

# supervisor_pattern.py
from langgraph.graph import StateGraph, END
from typing import TypedDict, Literal

class SupervisorState(TypedDict):
    messages: list
    next_agent: str
    task_status: dict

def supervisor(state: SupervisorState) -> SupervisorState:
    """Supervisor that delegates tasks to appropriate agents"""

    supervisor_prompt = """You are a supervisor managing a team of agents:
    - researcher: Finds and analyzes information
    - coder: Writes and reviews code
    - writer: Creates documentation and reports

    Based on the current task, decide which agent should handle it next.
    Respond with the agent name or 'FINISH' if the task is complete."""

    response = llm.invoke(state["messages"] + [HumanMessage(content=supervisor_prompt)])
    next_agent = response.content.strip().lower()

    return {
        "messages": state["messages"] + [response],
        "next_agent": next_agent
    }

def researcher(state: SupervisorState) -> SupervisorState:
    """Information gathering agent"""
    research_response = llm.invoke(
        state["messages"] +
        [HumanMessage(content="Research the topic and provide findings.")]
    )
    return {"messages": state["messages"] + [research_response]}

def coder(state: SupervisorState) -> SupervisorState:
    """Coding agent"""
    code_response = llm.invoke(
        state["messages"] +
        [HumanMessage(content="Write or review code for the task.")]
    )
    return {"messages": state["messages"] + [code_response]}

def writer(state: SupervisorState) -> SupervisorState:
    """Documentation agent"""
    write_response = llm.invoke(
        state["messages"] +
        [HumanMessage(content="Create documentation or a report.")]
    )
    return {"messages": state["messages"] + [write_response]}

def route_to_agent(state: SupervisorState) -> Literal["researcher", "coder", "writer", "end"]:
    next_agent = state["next_agent"]
    if next_agent == "finish":
        return "end"
    return next_agent

# Graph construction
supervisor_graph = StateGraph(SupervisorState)

supervisor_graph.add_node("supervisor", supervisor)
supervisor_graph.add_node("researcher", researcher)
supervisor_graph.add_node("coder", coder)
supervisor_graph.add_node("writer", writer)

supervisor_graph.set_entry_point("supervisor")

supervisor_graph.add_conditional_edges(
    "supervisor",
    route_to_agent,
    {
        "researcher": "researcher",
        "coder": "coder",
        "writer": "writer",
        "end": END
    }
)

# Return to supervisor after each agent completes
for agent in ["researcher", "coder", "writer"]:
    supervisor_graph.add_edge(agent, "supervisor")

multi_agent_app = supervisor_graph.compile()

---

7. Monitoring and Operations

Langfuse GenAI Observability

Langfuse es una plataforma de observabilidad para aplicaciones LLM.

# langfuse-deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: langfuse
  namespace: ai-monitoring
spec:
  replicas: 2
  selector:
    matchLabels:
      app: langfuse
  template:
    metadata:
      labels:
        app: langfuse
    spec:
      containers:
        - name: langfuse
          image: langfuse/langfuse:latest
          ports:
            - containerPort: 3000
          env:
            - name: DATABASE_URL
              valueFrom:
                secretKeyRef:
                  name: langfuse-secrets
                  key: database-url
            - name: NEXTAUTH_SECRET
              valueFrom:
                secretKeyRef:
                  name: langfuse-secrets
                  key: nextauth-secret
            - name: NEXTAUTH_URL
              value: "https://langfuse.example.com"
            - name: SALT
              valueFrom:
                secretKeyRef:
                  name: langfuse-secrets
                  key: salt
          resources:
            requests:
              cpu: "1"
              memory: "2Gi"
            limits:
              cpu: "2"
              memory: "4Gi"
---
apiVersion: v1
kind: Service
metadata:
  name: langfuse
  namespace: ai-monitoring
spec:
  selector:
    app: langfuse
  ports:
    - port: 3000
      targetPort: 3000
  type: ClusterIP

Langfuse Integration Code

# langfuse_integration.py
from langfuse import Langfuse
from langfuse.decorators import observe, langfuse_context
from openai import OpenAI

# Initialize Langfuse client
langfuse = Langfuse(
    public_key="pk-lf-xxx",
    secret_key="sk-lf-xxx",
    host="http://langfuse.ai-monitoring:3000"
)

client = OpenAI(
    api_key="sk-litellm-master-key",
    base_url="http://litellm-gateway.ai-gateway:4000/v1"
)

@observe()
def rag_query(user_query: str, user_id: str = None) -> str:
    """Track RAG queries with Langfuse"""

    # Set user ID
    langfuse_context.update_current_trace(
        user_id=user_id,
        tags=["rag", "production"]
    )

    # Document retrieval (tracked as separate span)
    with langfuse_context.observe(name="document_retrieval") as span:
        documents = search_documents(user_query)
        span.update(
            input={"query": user_query},
            output={"doc_count": len(documents)},
            metadata={"retrieval_method": "mmr"}
        )

    # LLM call
    with langfuse_context.observe(name="llm_generation") as span:
        response = client.chat.completions.create(
            model="llama3-70b",
            messages=[
                {"role": "system", "content": "Answer based on the context."},
                {"role": "user", "content": f"Context: {documents}\n\nQuestion: {user_query}"}
            ],
            max_tokens=1000
        )

        answer = response.choices[0].message.content

        # Track token usage and cost
        span.update(
            input={"messages": messages},
            output={"response": answer},
            usage={
                "input": response.usage.prompt_tokens,
                "output": response.usage.completion_tokens,
                "total": response.usage.total_tokens
            },
            metadata={
                "model": "llama3-70b",
                "temperature": 0.7
            }
        )

    return answer

# Collect feedback
def collect_feedback(trace_id: str, score: float, comment: str = None):
    """Record user feedback to Langfuse"""
    langfuse.score(
        trace_id=trace_id,
        name="user_feedback",
        value=score,
        comment=comment
    )

GPU Monitoring (DCGM)

# dcgm-exporter.yaml
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: dcgm-exporter
  namespace: gpu-monitoring
spec:
  selector:
    matchLabels:
      app: dcgm-exporter
  template:
    metadata:
      labels:
        app: dcgm-exporter
    spec:
      nodeSelector:
        nvidia.com/gpu.present: "true"
      tolerations:
        - key: nvidia.com/gpu
          operator: Exists
          effect: NoSchedule
      containers:
        - name: dcgm-exporter
          image: nvcr.io/nvidia/k8s/dcgm-exporter:3.3.5-3.4.0-ubuntu22.04
          ports:
            - containerPort: 9400
              name: metrics
          env:
            - name: DCGM_EXPORTER_LISTEN
              value: ":9400"
            - name: DCGM_EXPORTER_KUBERNETES
              value: "true"
          securityContext:
            privileged: true
          volumeMounts:
            - name: pod-resources
              mountPath: /var/lib/kubelet/pod-resources
      volumes:
        - name: pod-resources
          hostPath:
            path: /var/lib/kubelet/pod-resources
---
apiVersion: v1
kind: Service
metadata:
  name: dcgm-exporter
  namespace: gpu-monitoring
  labels:
    app: dcgm-exporter
spec:
  selector:
    app: dcgm-exporter
  ports:
    - port: 9400
      targetPort: 9400
      name: metrics
  clusterIP: None
---
apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
  name: dcgm-exporter
  namespace: gpu-monitoring
spec:
  selector:
    matchLabels:
      app: dcgm-exporter
  endpoints:
    - port: metrics
      interval: 15s

Key GPU Metrics

Metric	Description	Threshold
DCGM_FI_DEV_GPU_UTIL	GPU utilization	> 80% normal
DCGM_FI_DEV_MEM_COPY_UTIL	Memory bandwidth utilization	> 70% caution
DCGM_FI_DEV_FB_USED	Frame buffer usage	< 95% recommended
DCGM_FI_DEV_GPU_TEMP	GPU temperature	< 85C recommended
DCGM_FI_DEV_POWER_USAGE	Power consumption	Below 90% of TDP
DCGM_FI_DEV_SM_CLOCK	SM clock speed	Maintain default

Cost Optimization Strategies

1. Prompt Caching

# prompt_caching.py
import hashlib
import redis

redis_client = redis.Redis(host="redis.ai-cache", port=6379)

def get_cached_response(prompt: str, model: str) -> str | None:
    """Retrieve cached response"""
    cache_key = hashlib.sha256(f"{model}:{prompt}".encode()).hexdigest()
    cached = redis_client.get(cache_key)
    return cached.decode() if cached else None

def cache_response(prompt: str, model: str, response: str, ttl: int = 3600):
    """Cache response"""
    cache_key = hashlib.sha256(f"{model}:{prompt}".encode()).hexdigest()
    redis_client.setex(cache_key, ttl, response)

def query_with_cache(prompt: str, model: str = "llama3-70b") -> str:
    """Query with caching"""
    # Check cache
    cached = get_cached_response(prompt, model)
    if cached:
        return cached

    # LLM call
    response = client.chat.completions.create(
        model=model,
        messages=[{"role": "user", "content": prompt}]
    )
    result = response.choices[0].message.content

    # Cache result
    cache_response(prompt, model, result)
    return result

2. Tiered Model Selection

# tiered_model_selection.py
from enum import Enum

class TaskComplexity(Enum):
    SIMPLE = "simple"      # Classification, extraction, simple QA
    MODERATE = "moderate"  # Summarization, translation, general conversation
    COMPLEX = "complex"    # Analysis, reasoning, code generation

MODEL_TIERS = {
    TaskComplexity.SIMPLE: {
        "model": "llama3-8b",
        "cost_per_1k_tokens": 0.0001
    },
    TaskComplexity.MODERATE: {
        "model": "llama3-70b",
        "cost_per_1k_tokens": 0.0005
    },
    TaskComplexity.COMPLEX: {
        "model": "gpt-4o",
        "cost_per_1k_tokens": 0.01
    }
}

def classify_task_complexity(task: str) -> TaskComplexity:
    """Classify task complexity (using lightweight model)"""
    classification_prompt = f"""Classify the complexity of this task as SIMPLE, MODERATE, or COMPLEX:
    Task: {task}

    SIMPLE: Classification, extraction, simple QA
    MODERATE: Summarization, translation, general conversation
    COMPLEX: Analysis, reasoning, code generation

    Respond with only the classification."""

    response = client.chat.completions.create(
        model="llama3-8b",  # Use small model for classification
        messages=[{"role": "user", "content": classification_prompt}],
        max_tokens=10
    )

    classification = response.choices[0].message.content.strip().upper()
    return TaskComplexity[classification]

def execute_with_optimal_model(task: str) -> str:
    """Execute task with optimal model"""
    complexity = classify_task_complexity(task)
    model_config = MODEL_TIERS[complexity]

    response = client.chat.completions.create(
        model=model_config["model"],
        messages=[{"role": "user", "content": task}]
    )

    return response.choices[0].message.content

3. Batch Processing

# batch-processing-job.yaml
apiVersion: batch/v1
kind: CronJob
metadata:
  name: batch-inference
  namespace: ai-batch
spec:
  schedule: "0 2 * * *"  # Daily at 2 AM
  jobTemplate:
    spec:
      template:
        spec:
          containers:
            - name: batch-processor
              image: ai-platform/batch-processor:latest
              env:
                - name: BATCH_SIZE
                  value: "100"
                - name: MODEL
                  value: "llama3-70b"
                - name: QUEUE_URL
                  value: "redis://redis.ai-batch:6379/0"
              resources:
                requests:
                  cpu: "4"
                  memory: "8Gi"
          restartPolicy: OnFailure

4. Spot Instance Utilization

# spot-nodepool.yaml
apiVersion: karpenter.sh/v1
kind: NodePool
metadata:
  name: gpu-spot
spec:
  template:
    spec:
      requirements:
        - key: karpenter.sh/capacity-type
          operator: In
          values: ["spot"]
        - key: node.kubernetes.io/instance-type
          operator: In
          values:
            - g5.xlarge
            - g5.2xlarge
            - g6.xlarge
      taints:
        - key: spot-instance
          value: "true"
          effect: NoSchedule
  limits:
    nvidia.com/gpu: 50
  disruption:
    consolidationPolicy: WhenEmpty
    consolidateAfter: 1m

---

8. Evaluation and Quality Management

Ragas Framework

Ragas es un framework para evaluar la calidad de sistemas RAG.

# ragas_evaluation.py
from ragas import evaluate
from ragas.metrics import (
    faithfulness,
    answer_relevancy,
    context_precision,
    context_recall,
    answer_correctness
)
from datasets import Dataset

# Construct evaluation dataset
eval_data = {
    "question": [
        "What is a Kubernetes Pod?",
        "How does HPA work?"
    ],
    "answer": [
        "A Pod is the smallest deployable computing unit in Kubernetes.",
        "HPA automatically adjusts the number of Pods based on CPU utilization."
    ],
    "contexts": [
        ["A Pod is a group of one or more containers.", "Pods have shared storage and network."],
        ["HPA monitors metrics.", "It scales based on configured thresholds."]
    ],
    "ground_truth": [
        "A Pod is the smallest deployable computing unit that can be created and managed in Kubernetes.",
        "HPA automatically adjusts the number of workload replicas based on observed metrics (CPU, memory, etc.)."
    ]
}

dataset = Dataset.from_dict(eval_data)

# Run evaluation
results = evaluate(
    dataset,
    metrics=[
        faithfulness,        # Is the answer faithful to context
        answer_relevancy,    # Is the answer relevant to question
        context_precision,   # Is retrieved context precise
        context_recall,      # Was all necessary context retrieved
        answer_correctness   # Does answer match ground truth
    ]
)

print(results)
# {'faithfulness': 0.92, 'answer_relevancy': 0.88, 'context_precision': 0.85, ...}

Automated Evaluation Pipeline

# ragas-evaluation-cronjob.yaml
apiVersion: batch/v1
kind: CronJob
metadata:
  name: ragas-evaluation
  namespace: ai-qa
spec:
  schedule: "0 6 * * *"  # Daily at 6 AM
  jobTemplate:
    spec:
      template:
        spec:
          containers:
            - name: evaluator
              image: ai-platform/ragas-evaluator:latest
              env:
                - name: EVAL_DATASET_PATH
                  value: "s3://ai-datasets/eval/golden-set.json"
                - name: RAG_ENDPOINT
                  value: "http://rag-api.ai-inference:8000"
                - name: LANGFUSE_HOST
                  value: "http://langfuse.ai-monitoring:3000"
                - name: MIN_FAITHFULNESS
                  value: "0.85"
                - name: MIN_RELEVANCY
                  value: "0.80"
              resources:
                requests:
                  cpu: "2"
                  memory: "4Gi"
          restartPolicy: OnFailure

A/B Testing

# ab-testing-config.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: ab-testing-config
  namespace: ai-inference
data:
  config.yaml: |
    experiments:
      - name: llama3-70b-vs-gpt4o
        traffic_split:
          variant_a:
            model: llama3-70b
            weight: 80
          variant_b:
            model: gpt-4o
            weight: 20
        metrics:
          - latency_p99
          - user_satisfaction
          - cost_per_query
        duration_days: 14

      - name: chunk-size-experiment
        traffic_split:
          variant_a:
            chunk_size: 512
            weight: 50
          variant_b:
            chunk_size: 1024
            weight: 50
        metrics:
          - context_precision
          - answer_relevancy
        duration_days: 7

---

9. Core Technology Stack Summary

Technology	Purpose	Key Features
Kagent	AI Agent Lifecycle	CRD-based agent management, auto-scaling
Kgateway	Inference Gateway	InferencePool, Prefix-aware routing
Milvus	Vector Database	Large-scale vector search, GPU-accelerated indexing
Ragas	RAG Evaluation	Faithfulness, relevancy, accuracy metrics
LiteLLM	LLM Integrated Gateway	Provider abstraction, fallback, cost tracking
LangGraph	Workflow Orchestration	State management, conditional branching, error handling
Langfuse	GenAI Observability	Request tracing, cost analysis, feedback collection
vLLM	High-Performance Inference	PagedAttention, continuous batching, prefix caching
Karpenter	Node Provisioning	GPU node auto-scaling, Spot management
DCGM	GPU Monitoring	Utilization, temperature, power metrics

---

10. Next Steps

Practice Quiz

Para verificar tu comprensión de la plataforma de Agentic AI, realiza el siguiente quiz:

• Quiz de la plataforma de Agentic AI

Related Documents

• Guía detallada de despliegue de vLLM - Instalación y optimización detalladas de vLLM
• Cargas de trabajo AI/ML - Gestión de cargas de trabajo AI/ML en Kubernetes

References

• AI on EKS - Guía y ejemplos de AWS para desplegar cargas de trabajo AI/ML en EKS
• Documentación oficial de vLLM
• Documentación de LangGraph
• Documentación de Milvus
• Documentación de Langfuse
• NVIDIA GPU Operator
• Gateway API for AI