AWS CLI 또는 SDK를 사용하여 인스턴스 시작하기

aws bedrock-agentcore-control create-capacity-provider \
  --name "my_capacity_provider" \
  --permissions-configuration '{
    "capacityProviderOperatorRoleArn": "arn:aws:iam::111122223333:role/AgentCoreCapacityProviderOperatorRole"
  }' \
  --compute-configuration '{
    "ec2Configuration": {
      "launchTemplateSource": {
        "launchParameters": {
          "operatingSystem": "LINUX_X86_64",
          "instanceRequirements": {
            "allowedInstanceTypes": ["m5.large"]
          }
        }
      },
      "vpcConfiguration": {
        "subnets": ["subnet-0123456789abcdef0"],
        "securityGroups": ["sg-0123456789abcdef0"]
      },
      "lifecycleConfiguration": {
        "maxLifetime": 3600
      },
      "volumes": [
        { "ebsConfiguration": { "name": "scratch", "sizeGiB": 50, "volumeType": "gp3" } }
      ]
    }
  }'

boto3를 사용하여 용량 공급자를 생성하는 Python 예제입니다.

import boto3

client = boto3.client("bedrock-agentcore-control", region_name="us-west-2")

response = client.create_capacity_provider(
    name="my_capacity_provider",
    permissionsConfiguration={
        "capacityProviderOperatorRoleArn": "arn:aws:iam::111122223333:role/AgentCoreCapacityProviderOperatorRole"
    },
    computeConfiguration={
        "ec2Configuration": {
            "launchTemplateSource": {
                "launchParameters": {
                    "operatingSystem": "LINUX_X86_64",
                    "instanceRequirements": {
                        "allowedInstanceTypes": ["m5.large"],
                    },
                }
            },
            "vpcConfiguration": {
                "subnets": ["subnet-0123456789abcdef0"],
                "securityGroups": ["sg-0123456789abcdef0"],
            },
            "lifecycleConfiguration": {
                "maxLifetime": 3600,
            },
            "volumes": [
                {"ebsConfiguration": {"name": "scratch", "sizeGiB": 50, "volumeType": "gp3"}}
            ],
        }
    },
)

print(f"Capacity provider ARN: {response['capacityProviderArn']}")

aws bedrock-agentcore-control create-agent-runtime \
  --agent-runtime-name "my_instances_agent" \
  --role-arn "arn:aws:iam::111122223333:role/AgentRuntimeRole" \
  --agent-runtime-artifact '{
    "containerConfiguration": {
      "containerUri": "111122223333.dkr.ecr.us-west-2.amazonaws.com/my-agent:latest"
    }
  }' \
  --capacity-provider-configuration '{
    "capacityProviderArn": "arn:aws:bedrock-agentcore:us-west-2:111122223333:capacity-provider/my_capacity_provider-a1b2c3d4e5"
  }' \
  --lifecycle-configuration '{
    "idleRuntimeSessionTimeout": 300,
    "maxLifetime": 1800
  }' \
  --filesystem-configurations '[
    {
      "capacityProviderVolume": { "volumeName": "scratch", "mountPath": "/mnt/scratch" }
    }
  ]'

boto3를 사용하여 인스턴스에서 에이전트 런타임을 생성하는 Python 예제입니다.

import boto3

client = boto3.client("bedrock-agentcore-control", region_name="us-west-2")

response = client.create_agent_runtime(
    agentRuntimeName="my_instances_agent",
    roleArn="arn:aws:iam::111122223333:role/AgentRuntimeRole",
    agentRuntimeArtifact={
        "containerConfiguration": {
            "containerUri": "111122223333.dkr.ecr.us-west-2.amazonaws.com/my-agent:latest"
        }
    },
    capacityProviderConfiguration={
        "capacityProviderArn": "arn:aws:bedrock-agentcore:us-west-2:111122223333:capacity-provider/my_capacity_provider-a1b2c3d4e5",
    },
    lifecycleConfiguration={
        "idleRuntimeSessionTimeout": 300,
        "maxLifetime": 1800,
    },
    filesystemConfigurations=[
        {
            "capacityProviderVolume": {"volumeName": "scratch", "mountPath": "/mnt/scratch"}
        }
    ],
)

print(f"Agent runtime ARN: {response['agentRuntimeArn']}")

echo '{"prompt": "Analyze the sales data and summarize the key trends."}' > payload.json

aws bedrock-agentcore invoke-agent-runtime \
  --agent-runtime-arn "arn:aws:bedrock-agentcore:us-west-2:111122223333:runtime/my_instances_agent-suffix" \
  --runtime-session-id "project-xyz-0000000000000000000000000000" \
  --qualifier "DEFAULT" \
  --payload fileb://payload.json \
  response.json

payload 파라미터는 바이너리 Blob이며 AWS CLI는 기본적으로 Blob이 base64로 인코딩될 것으로 예상합니다. 와 같은 원시 JSON 인라인 전달은 요청에 도달하기 전에 Invalid base64 오류와 함께 클라이언트 측에서 --payload '{"prompt": "…​"}'실패합니다. 여기에 사용된 fileb:// 양식은 cli_binary_format 설정에 관계없이 파일을 바이너리로 읽으므로 인코딩 단계를 방지합니다.

aws bedrock-agentcore stop-runtime-session \
  --agent-runtime-arn "arn:aws:bedrock-agentcore:us-west-2:111122223333:runtime/my_instances_agent-suffix" \
  --runtime-session-id "project-xyz-0000000000000000000000000000"

boto3를 사용하여 런타임 세션을 중지하는 Python 예제입니다.

import boto3

client = boto3.client("bedrock-agentcore", region_name="us-west-2")

client.stop_runtime_session(
    agentRuntimeArn="arn:aws:bedrock-agentcore:us-west-2:111122223333:runtime/my_instances_agent-suffix",
    runtimeSessionId="project-xyz-0000000000000000000000000000",
)

aws bedrock-agentcore delete-capacity-provider-session \
  --capacity-provider-id "my_capacity_provider-a1b2c3d4e5" \
  --session-id "project-xyz-0000000000000000000000000000"

boto3를 사용하여 용량 공급자 세션을 삭제하는 Python 예제입니다.

import boto3

client = boto3.client("bedrock-agentcore", region_name="us-west-2")

response = client.delete_capacity_provider_session(
    capacityProviderId="my_capacity_provider-a1b2c3d4e5",
    sessionId="project-xyz-0000000000000000000000000000",
)

print("Session status:", response["status"])

aws bedrock-agentcore-control delete-capacity-provider \
  --capacity-provider-id "my_capacity_provider-a1b2c3d4e5"

boto3를 사용하여 용량 공급자를 삭제하는 Python 예제입니다.

import boto3

client = boto3.client("bedrock-agentcore-control", region_name="us-west-2")

client.delete_capacity_provider(
    capacityProviderId="my_capacity_provider-a1b2c3d4e5",
)