Back to Details

bedrock-agentcore-deployment

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Amazon Bedrock AgentCore Deployment

Amazon Bedrock AgentCore 部署方案

Overview

概述

Deploy AI agents to Amazon Bedrock AgentCore using multiple approaches: starter toolkit for rapid deployment, direct code deployment for customization, and container deployment for complex dependencies. Includes CI/CD patterns and infrastructure as code.
Purpose: Deploy agents from development to production with best practices
Pattern: Workflow-based (4 deployment methods)
Key Principles (validated by AWS December 2025):
  1. Zero Infrastructure - Managed compute, no servers to manage
  2. arm64 Architecture - All deployments use arm64
  3. Session Isolation - Complete isolation between sessions
  4. Multiple Entry Points - SDK decorator or REST endpoints
  5. Observability Built-in - CloudWatch and OTel integration
  6. Framework Agnostic - Works with any Python agent framework
Quality Targets:
  • Deployment time: < 5 minutes
  • Cold start: < 2 seconds
  • Package size: 250MB (zip), 750MB (unzipped)
通过多种方式将AI Agent部署到Amazon Bedrock AgentCore：用于快速部署的入门工具包、用于自定义的直接代码部署，以及用于复杂依赖项的容器部署。同时包含CI/CD模式和基础设施即代码方案。
用途：遵循最佳实践将Agent从开发环境部署到生产环境
模式：基于工作流（4种部署方法）
核心原则（经AWS 2025年12月验证）：
  1. 零基础设施管理 - 托管式计算，无需管理服务器
  2. arm64架构 - 所有部署均采用arm64
  3. 会话隔离 - 会话之间完全隔离
  4. 多入口点 - SDK装饰器或REST端点
  5. 内置可观测性 - 集成CloudWatch和OTel
  6. 框架无关 - 兼容任意Python Agent框架
质量指标：
  • 部署时间：< 5分钟
  • 冷启动时间：< 2秒
  • 包大小：250MB（压缩包），750MB（解压后）

When to Use

适用场景

Use bedrock-agentcore-deployment when:
  • Deploying agent code to production
  • Setting up CI/CD pipelines for agents
  • Migrating from development to production
  • Managing multiple agent versions
  • Implementing blue-green deployments
When NOT to Use:
  • Local development/testing (use local server)
  • Standard Bedrock Agents with action groups
在以下场景中使用bedrock-agentcore-deployment：
  • 将Agent代码部署到生产环境
  • 为Agent搭建CI/CD流水线
  • 从开发环境迁移到生产环境
  • 管理多个Agent版本
  • 实现蓝绿部署
不适用场景：
  • 本地开发/测试（使用本地服务器）
  • 带操作组的标准Bedrock Agents

Prerequisites

前置条件

Required

必需项

  • AWS account with AgentCore access
  • Python 3.10+ (recommended: 3.13)
  • IAM role with AgentCore permissions
  • Foundation model access enabled
  • 拥有AgentCore访问权限的AWS账户
  • Python 3.10+（推荐版本：3.13）
  • 拥有AgentCore权限的IAM角色
  • 已启用基础模型访问权限

Recommended

推荐项

  • AWS CLI configured
  • Docker (for container deployments)
  • GitHub Actions or GitLab CI (for CI/CD)
  • 已配置AWS CLI
  • Docker（用于容器部署）
  • GitHub Actions或GitLab CI（用于CI/CD）

Deployment Method 1: Starter Toolkit (Fastest)

部署方法1：入门工具包（最快）

Time: 2-5 minutes Complexity: Low Best For: Rapid deployment, simple agents
耗时：2-5分钟 复杂度：低 最佳适用：快速部署、简单Agent

Step 1: Install Toolkit

步骤1：安装工具包

bash
pip install bedrock-agentcore strands-agents bedrock-agentcore-starter-toolkit
bash
pip install bedrock-agentcore strands-agents bedrock-agentcore-starter-toolkit

Verify

验证

agentcore --help
undefined
agentcore --help
undefined

Step 2: Create Agent

步骤2：创建Agent

python
undefined
python
undefined

main.py

main.py

from bedrock_agentcore import BedrockAgentCoreApp from strands import Agent
app = BedrockAgentCoreApp() agent = Agent(model="anthropic.claude-sonnet-4-20250514-v1:0")
@app.entrypoint def invoke(payload): prompt = payload.get("prompt", "Hello!") result = agent(prompt) return {"response": result.message}
if name == "main": app.run()
undefined
from bedrock_agentcore import BedrockAgentCoreApp from strands import Agent
app = BedrockAgentCoreApp() agent = Agent(model="anthropic.claude-sonnet-4-20250514-v1:0")
@app.entrypoint def invoke(payload): prompt = payload.get("prompt", "Hello!") result = agent(prompt) return {"response": result.message}
if name == "main": app.run()
undefined

Step 3: Configure

步骤3：配置

bash
undefined
bash
undefined

Initialize configuration

初始化配置

agentcore configure -e main.py -n my-production-agent
agentcore configure -e main.py -n my-production-agent

This creates .bedrock_agentcore.yaml

此命令会生成.bedrock_agentcore.yaml文件

undefined
undefined

Step 4: Test Locally

步骤4：本地测试

bash
undefined
bash
undefined

Start local server

启动本地服务器

python main.py &
python main.py &

Test

测试

curl -X POST http://localhost:8080/invocations
-H "Content-Type: application/json"
-d '{"prompt": "Hello, world!"}'
curl -X POST http://localhost:8080/invocations
-H "Content-Type: application/json"
-d '{"prompt": "Hello, world!"}'

Stop local server

停止本地服务器

pkill -f main.py
undefined
pkill -f main.py
undefined

Step 5: Deploy

步骤5：部署到AWS

bash
undefined
bash
undefined

Deploy to AWS

部署到AWS

agentcore deploy
agentcore deploy

Output: Agent ARN

输出：Agent ARN

arn:aws:bedrock-agentcore:us-east-1:123456789012:agent-runtime/my-production-agent

arn:aws:bedrock-agentcore:us-east-1:123456789012:agent-runtime/my-production-agent

undefined
undefined

Step 6: Test Deployed

步骤6：测试已部署的Agent

bash
undefined
bash
undefined

Test via CLI

通过CLI测试

agentcore invoke '{"prompt": "Hello from production!"}'
agentcore invoke '{"prompt": "Hello from production!"}'

Test via boto3

通过boto3测试

python -c " import boto3 client = boto3.client('bedrock-agentcore') response = client.invoke_agent_runtime( agentRuntimeArn='arn:...', runtimeSessionId='test-1', payload={'prompt': 'Hello!'} ) print(response['payload']) "

---
python -c " import boto3 client = boto3.client('bedrock-agentcore') response = client.invoke_agent_runtime( agentRuntimeArn='arn:...', runtimeSessionId='test-1', payload={'prompt': 'Hello!'} ) print(response['payload']) "

---

Deployment Method 2: Direct Code Deploy

部署方法2：直接代码部署

Time: 10-15 minutes Complexity: Medium Best For: Custom dependencies, specific configurations
耗时：10-15分钟 复杂度：中 最佳适用：自定义依赖项、特定配置

Step 1: Project Structure

步骤1：项目结构

my-agent/
├── main.py              # Entry point
├── agent/
│   ├── __init__.py
│   └── logic.py         # Agent logic
├── pyproject.toml       # Dependencies
└── requirements.txt     # Alternative deps format
my-agent/
├── main.py              # 入口文件
├── agent/
│   ├── __init__.py
│   └── logic.py         # Agent逻辑代码
├── pyproject.toml       # 依赖配置
└── requirements.txt     # 依赖配置的替代格式

Step 2: Create Entry Point

步骤2：创建入口文件

python
undefined
python
undefined

main.py - REST endpoint pattern

main.py - REST端点模式

from flask import Flask, request, jsonify
app = Flask(name)
@app.route('/invocations', methods=['POST']) def invoke(): payload = request.get_json() prompt = payload.get('prompt', '')
# Your agent logic here
from agent.logic import process_request
result = process_request(prompt)

return jsonify({'response': result})
@app.route('/ping', methods=['GET']) def ping(): return 'OK', 200
if name == 'main': app.run(host='0.0.0.0', port=8080)
undefined
from flask import Flask, request, jsonify
app = Flask(name)
@app.route('/invocations', methods=['POST']) def invoke(): payload = request.get_json() prompt = payload.get('prompt', '')
# 在此处添加你的Agent逻辑
from agent.logic import process_request
result = process_request(prompt)

return jsonify({'response': result})
@app.route('/ping', methods=['GET']) def ping(): return 'OK', 200
if name == 'main': app.run(host='0.0.0.0', port=8080)
undefined

Step 3: Package for arm64

步骤3：为arm64架构打包

bash
undefined
bash
undefined

Create virtual environment

创建虚拟环境

uv init agent-deploy --python 3.13 cd agent-deploy
uv init agent-deploy --python 3.13 cd agent-deploy

Install dependencies for arm64

为arm64安装依赖

uv pip install
--python-platform aarch64-manylinux2014
--python-version 3.13
--target=deployment_package
--only-binary=:all:
-r requirements.txt
uv pip install
--python-platform aarch64-manylinux2014
--python-version 3.13
--target=deployment_package
--only-binary=:all:
-r requirements.txt

Create ZIP

创建ZIP包

cd deployment_package zip -r ../deployment_package.zip . cd ..
cd deployment_package zip -r ../deployment_package.zip . cd ..

Add main.py

添加main.py

zip deployment_package.zip main.py
zip deployment_package.zip main.py

Add agent module

添加agent模块

zip -r deployment_package.zip agent/
undefined
zip -r deployment_package.zip agent/
undefined

Step 4: Upload to S3

步骤4：上传到S3

bash
aws s3 cp deployment_package.zip s3://my-bucket/agents/v1.0.0/package.zip
bash
aws s3 cp deployment_package.zip s3://my-bucket/agents/v1.0.0/package.zip

Step 5: Create Agent Runtime

步骤5：创建Agent Runtime

python
import boto3

control = boto3.client('bedrock-agentcore-control')

response = control.create_agent_runtime(
    name='my-custom-agent',
    description='Production agent with custom dependencies',
    agentRuntimeArtifact={
        's3': {
            'uri': 's3://my-bucket/agents/v1.0.0/package.zip'
        }
    },
    roleArn='arn:aws:iam::123456789012:role/AgentCoreExecutionRole',
    pythonRuntime='PYTHON_3_13',
    entryPoint=['main.py'],
    environmentVariables={
        'LOG_LEVEL': 'INFO',
        'CUSTOM_CONFIG': 'production'
    }
)

agent_arn = response['agentRuntimeArn']
print(f"Deployed: {agent_arn}")
python
import boto3

control = boto3.client('bedrock-agentcore-control')

response = control.create_agent_runtime(
    name='my-custom-agent',
    description='Production agent with custom dependencies',
    agentRuntimeArtifact={
        's3': {
            'uri': 's3://my-bucket/agents/v1.0.0/package.zip'
        }
    },
    roleArn='arn:aws:iam::123456789012:role/AgentCoreExecutionRole',
    pythonRuntime='PYTHON_3_13',
    entryPoint=['main.py'],
    environmentVariables={
        'LOG_LEVEL': 'INFO',
        'CUSTOM_CONFIG': 'production'
    }
)

agent_arn = response['agentRuntimeArn']
print(f"Deployed: {agent_arn}")

Deployment Method 3: Container Deploy

部署方法3：容器部署

Time: 15-30 minutes Complexity: High Best For: Complex dependencies, custom runtimes, large packages
耗时：15-30分钟 复杂度：高 最佳适用：复杂依赖项、自定义运行时、大型包

Step 1: Create Dockerfile

步骤1：创建Dockerfile

dockerfile
undefined
dockerfile
undefined

Dockerfile

Dockerfile

FROM public.ecr.aws/lambda/python:3.13-arm64
FROM public.ecr.aws/lambda/python:3.13-arm64

Install dependencies

安装依赖

COPY requirements.txt . RUN pip install --no-cache-dir -r requirements.txt
COPY requirements.txt . RUN pip install --no-cache-dir -r requirements.txt

Copy agent code

复制Agent代码

COPY main.py . COPY agent/ agent/
COPY main.py . COPY agent/ agent/

Set entry point

设置入口点

ENV PORT=8080 EXPOSE 8080
CMD ["main.py"]
undefined
ENV PORT=8080 EXPOSE 8080
CMD ["main.py"]
undefined

Step 2: Build Image

步骤2：构建镜像

bash
undefined
bash
undefined

Login to ECR

登录ECR

aws ecr get-login-password --region us-east-1 |
docker login --username AWS --password-stdin 123456789012.dkr.ecr.us-east-1.amazonaws.com
aws ecr get-login-password --region us-east-1 |
docker login --username AWS --password-stdin 123456789012.dkr.ecr.us-east-1.amazonaws.com

Build for arm64

为arm64架构构建

docker buildx build
--platform linux/arm64
-t 123456789012.dkr.ecr.us-east-1.amazonaws.com/my-agent:v1.0.0
--push .
undefined
docker buildx build
--platform linux/arm64
-t 123456789012.dkr.ecr.us-east-1.amazonaws.com/my-agent:v1.0.0
--push .
undefined

Step 3: Create Agent Runtime

步骤3：创建Agent Runtime

python
response = control.create_agent_runtime(
    name='my-container-agent',
    description='Agent deployed via container',
    agentRuntimeArtifact={
        'container': {
            'imageUri': '123456789012.dkr.ecr.us-east-1.amazonaws.com/my-agent:v1.0.0'
        }
    },
    roleArn='arn:aws:iam::123456789012:role/AgentCoreExecutionRole'
)
python
response = control.create_agent_runtime(
    name='my-container-agent',
    description='Agent deployed via container',
    agentRuntimeArtifact={
        'container': {
            'imageUri': '123456789012.dkr.ecr.us-east-1.amazonaws.com/my-agent:v1.0.0'
        }
    },
    roleArn='arn:aws:iam::123456789012:role/AgentCoreExecutionRole'
)

Deployment Method 4: Infrastructure as Code (Terraform)

部署方法4：基础设施即代码（Terraform）

Time: 20-30 minutes setup, then automated Complexity: High Best For: Production environments, team deployments
耗时：20-30分钟配置，之后可自动化部署 复杂度：高 最佳适用：生产环境、团队协作部署

Terraform Configuration

Terraform配置

hcl
undefined
hcl
undefined

main.tf

main.tf

terraform { required_providers { aws = { source = "hashicorp/aws" version = "~> 5.0" } } }
provider "aws" { region = "us-east-1" }
terraform { required_providers { aws = { source = "hashicorp/aws" version = "~> 5.0" } } }
provider "aws" { region = "us-east-1" }

IAM Role for Agent

Agent的IAM角色

resource "aws_iam_role" "agentcore_execution" { name = "agentcore-execution-role"
assume_role_policy = jsonencode({ Version = "2012-10-17" Statement = [{ Action = "sts:AssumeRole" Effect = "Allow" Principal = { Service = "bedrock-agentcore.amazonaws.com" } }] }) }
resource "aws_iam_role_policy" "agentcore_policy" { name = "agentcore-policy" role = aws_iam_role.agentcore_execution.id
policy = jsonencode({ Version = "2012-10-17" Statement = [ { Effect = "Allow" Action = [ "bedrock:InvokeModel", "bedrock:InvokeModelWithResponseStream" ] Resource = "arn:aws:bedrock:::foundation-model/" }, { Effect = "Allow" Action = [ "logs:CreateLogGroup", "logs:CreateLogStream", "logs:PutLogEvents" ] Resource = "" }, { Effect = "Allow" Action = [ "s3:GetObject" ] Resource = "${aws_s3_bucket.agent_artifacts.arn}/" } ] }) }
resource "aws_iam_role" "agentcore_execution" { name = "agentcore-execution-role"
assume_role_policy = jsonencode({ Version = "2012-10-17" Statement = [{ Action = "sts:AssumeRole" Effect = "Allow" Principal = { Service = "bedrock-agentcore.amazonaws.com" } }] }) }
resource "aws_iam_role_policy" "agentcore_policy" { name = "agentcore-policy" role = aws_iam_role.agentcore_execution.id
policy = jsonencode({ Version = "2012-10-17" Statement = [ { Effect = "Allow" Action = [ "bedrock:InvokeModel", "bedrock:InvokeModelWithResponseStream" ] Resource = "arn:aws:bedrock:::foundation-model/" }, { Effect = "Allow" Action = [ "logs:CreateLogGroup", "logs:CreateLogStream", "logs:PutLogEvents" ] Resource = "" }, { Effect = "Allow" Action = [ "s3:GetObject" ] Resource = "${aws_s3_bucket.agent_artifacts.arn}/" } ] }) }

S3 Bucket for Agent Artifacts

用于存储Agent制品的S3桶

resource "aws_s3_bucket" "agent_artifacts" { bucket = "my-agentcore-artifacts-${data.aws_caller_identity.current.account_id}" }
resource "aws_s3_bucket" "agent_artifacts" { bucket = "my-agentcore-artifacts-${data.aws_caller_identity.current.account_id}" }

Upload agent package

上传Agent包

resource "aws_s3_object" "agent_package" { bucket = aws_s3_bucket.agent_artifacts.id key = "agents/${var.agent_version}/package.zip" source = "${path.module}/deployment_package.zip" etag = filemd5("${path.module}/deployment_package.zip") }
resource "aws_s3_object" "agent_package" { bucket = aws_s3_bucket.agent_artifacts.id key = "agents/${var.agent_version}/package.zip" source = "${path.module}/deployment_package.zip" etag = filemd5("${path.module}/deployment_package.zip") }

Note: AgentCore resources may require custom provider or AWS CLI

注意：AgentCore资源可能需要自定义Provider或AWS CLI

Use null_resource with local-exec as workaround

使用null_resource结合local-exec作为临时解决方案

resource "null_resource" "create_agent_runtime" { triggers = { package_etag = aws_s3_object.agent_package.etag }
provisioner "local-exec" { command = <<-EOT aws bedrock-agentcore-control create-agent-runtime
--name ${var.agent_name}
--description "Terraform-managed agent"
--agent-runtime-artifact '{"s3":{"uri":"s3://${aws_s3_bucket.agent_artifacts.id}/agents/${var.agent_version}/package.zip"}}'
--role-arn ${aws_iam_role.agentcore_execution.arn}
--python-runtime PYTHON_3_13
--entry-point '["main.py"]' EOT }
depends_on = [aws_s3_object.agent_package] }
data "aws_caller_identity" "current" {}
variable "agent_name" { default = "my-production-agent" }
variable "agent_version" { default = "1.0.0" }

---
resource "null_resource" "create_agent_runtime" { triggers = { package_etag = aws_s3_object.agent_package.etag }
provisioner "local-exec" { command = <<-EOT aws bedrock-agentcore-control create-agent-runtime
--name ${var.agent_name}
--description "Terraform-managed agent"
--agent-runtime-artifact '{"s3":{"uri":"s3://${aws_s3_bucket.agent_artifacts.id}/agents/${var.agent_version}/package.zip"}}'
--role-arn ${aws_iam_role.agentcore_execution.arn}
--python-runtime PYTHON_3_13
--entry-point '["main.py"]' EOT }
depends_on = [aws_s3_object.agent_package] }
data "aws_caller_identity" "current" {}
variable "agent_name" { default = "my-production-agent" }
variable "agent_version" { default = "1.0.0" }

---

CI/CD Pipeline: GitHub Actions

CI/CD流水线：GitHub Actions

yaml
undefined
yaml
undefined

.github/workflows/deploy-agent.yml

.github/workflows/deploy-agent.yml

name: Deploy AgentCore Agent
on: push: branches: [main] pull_request: branches: [main]
env: AWS_REGION: us-east-1 AGENT_NAME: my-production-agent
jobs: test: runs-on: ubuntu-latest steps: - uses: actions/checkout@v4
  - name: Set up Python
    uses: actions/setup-python@v5
    with:
      python-version: '3.13'

  - name: Install dependencies
    run: |
      pip install -r requirements.txt
      pip install pytest

  - name: Run tests
    run: pytest tests/

  - name: Test local server
    run: |
      python main.py &
      sleep 5
      curl -f http://localhost:8080/ping
      curl -X POST http://localhost:8080/invocations \
        -H "Content-Type: application/json" \
        -d '{"prompt": "test"}'
      pkill -f main.py
deploy: needs: test if: github.ref == 'refs/heads/main' runs-on: ubuntu-latest permissions: id-token: write contents: read
steps:
  - uses: actions/checkout@v4

  - name: Configure AWS credentials
    uses: aws-actions/configure-aws-credentials@v4
    with:
      role-to-assume: arn:aws:iam::${{ secrets.AWS_ACCOUNT_ID }}:role/GitHubActionsRole
      aws-region: ${{ env.AWS_REGION }}

  - name: Install AgentCore toolkit
    run: pip install bedrock-agentcore-starter-toolkit

  - name: Configure agent
    run: agentcore configure -e main.py -n ${{ env.AGENT_NAME }}

  - name: Deploy agent
    run: agentcore deploy

  - name: Verify deployment
    run: |
      agentcore invoke '{"prompt": "Health check"}'

---
name: Deploy AgentCore Agent
on: push: branches: [main] pull_request: branches: [main]
env: AWS_REGION: us-east-1 AGENT_NAME: my-production-agent
jobs: test: runs-on: ubuntu-latest steps: - uses: actions/checkout@v4
  - name: Set up Python
    uses: actions/setup-python@v5
    with:
      python-version: '3.13'

  - name: Install dependencies
    run: |
      pip install -r requirements.txt
      pip install pytest

  - name: Run tests
    run: pytest tests/

  - name: Test local server
    run: |
      python main.py &
      sleep 5
      curl -f http://localhost:8080/ping
      curl -X POST http://localhost:8080/invocations \
        -H "Content-Type: application/json" \
        -d '{"prompt": "test"}'
      pkill -f main.py
deploy: needs: test if: github.ref == 'refs/heads/main' runs-on: ubuntu-latest permissions: id-token: write contents: read
steps:
  - uses: actions/checkout@v4

  - name: Configure AWS credentials
    uses: aws-actions/configure-aws-credentials@v4
    with:
      role-to-assume: arn:aws:iam::${{ secrets.AWS_ACCOUNT_ID }}:role/GitHubActionsRole
      aws-region: ${{ env.AWS_REGION }}

  - name: Install AgentCore toolkit
    run: pip install bedrock-agentcore-starter-toolkit

  - name: Configure agent
    run: agentcore configure -e main.py -n ${{ env.AGENT_NAME }}

  - name: Deploy agent
    run: agentcore deploy

  - name: Verify deployment
    run: |
      agentcore invoke '{"prompt": "Health check"}'

---

CI/CD Pipeline: GitLab CI

CI/CD流水线：GitLab CI

yaml
undefined
yaml
undefined

.gitlab-ci.yml

.gitlab-ci.yml

stages:
  • test
  • build
  • deploy
variables: AWS_REGION: us-east-1 AGENT_NAME: my-production-agent
test: stage: test image: python:3.13 script: - pip install -r requirements.txt - pip install pytest - pytest tests/
build: stage: build image: python:3.13 script: - pip install uv - uv pip install --python-platform aarch64-manylinux2014 --python-version 3.13 --target=deployment_package --only-binary=:all: -r requirements.txt - cd deployment_package && zip -r ../package.zip . && cd .. - zip package.zip main.py artifacts: paths: - package.zip
deploy: stage: deploy image: amazon/aws-cli script: - aws s3 cp package.zip s3://${ARTIFACT_BUCKET}/agents/${CI_COMMIT_SHA}/package.zip - | aws bedrock-agentcore-control update-agent-runtime
--agent-runtime-id ${AGENT_RUNTIME_ID}
--agent-runtime-artifact "{"s3":{"uri":"s3://${ARTIFACT_BUCKET}/agents/${CI_COMMIT_SHA}/package.zip"}}" only: - main

---
stages:
  • test
  • build
  • deploy
variables: AWS_REGION: us-east-1 AGENT_NAME: my-production-agent
test: stage: test image: python:3.13 script: - pip install -r requirements.txt - pip install pytest - pytest tests/
build: stage: build image: python:3.13 script: - pip install uv - uv pip install --python-platform aarch64-manylinux2014 --python-version 3.13 --target=deployment_package --only-binary=:all: -r requirements.txt - cd deployment_package && zip -r ../package.zip . && cd .. - zip package.zip main.py artifacts: paths: - package.zip
deploy: stage: deploy image: amazon/aws-cli script: - aws s3 cp package.zip s3://${ARTIFACT_BUCKET}/agents/${CI_COMMIT_SHA}/package.zip - | aws bedrock-agentcore-control update-agent-runtime
--agent-runtime-id ${AGENT_RUNTIME_ID}
--agent-runtime-artifact "{"s3":{"uri":"s3://${ARTIFACT_BUCKET}/agents/${CI_COMMIT_SHA}/package.zip"}}" only: - main

---

Version Management

版本管理

Blue-Green Deployment

蓝绿部署

python
def blue_green_deploy(agent_name, new_version):
    """Deploy new version alongside old, then switch"""

    # Get current (blue) version
    blue = control.get_agent_runtime(name=f"{agent_name}-blue")

    # Deploy new (green) version
    green = control.create_agent_runtime(
        name=f"{agent_name}-green",
        agentRuntimeArtifact={'s3': {'uri': f's3://bucket/agents/{new_version}/package.zip'}},
        roleArn=blue['roleArn'],
        pythonRuntime='PYTHON_3_13',
        entryPoint=['main.py']
    )

    # Test green
    test_result = run_smoke_tests(green['agentRuntimeArn'])

    if test_result.passed:
        # Update endpoint to point to green
        control.update_agent_runtime_endpoint(
            endpointId='production-endpoint',
            agentRuntimeArn=green['agentRuntimeArn']
        )
        # Delete old blue
        control.delete_agent_runtime(name=f"{agent_name}-blue")
        # Rename green to blue
        # (Note: actual rename may require recreate)
    else:
        # Rollback - delete failed green
        control.delete_agent_runtime(name=f"{agent_name}-green")
        raise Exception("Green deployment failed tests")
python
def blue_green_deploy(agent_name, new_version):
    """在旧版本旁部署新版本，然后切换流量"""

    # 获取当前（蓝色）版本
    blue = control.get_agent_runtime(name=f"{agent_name}-blue")

    # 部署新（绿色）版本
    green = control.create_agent_runtime(
        name=f"{agent_name}-green",
        agentRuntimeArtifact={'s3': {'uri': f's3://bucket/agents/${new_version}/package.zip'}},
        roleArn=blue['roleArn'],
        pythonRuntime='PYTHON_3_13',
        entryPoint=['main.py']
    )

    # 测试绿色版本
    test_result = run_smoke_tests(green['agentRuntimeArn'])

    if test_result.passed:
        # 更新端点指向绿色版本
        control.update_agent_runtime_endpoint(
            endpointId='production-endpoint',
            agentRuntimeArn=green['agentRuntimeArn']
        )
        # 删除旧的蓝色版本
        control.delete_agent_runtime(name=f"{agent_name}-blue")
        # 将绿色版本重命名为蓝色版本
        # （注意：实际重命名可能需要重新创建）
    else:
        # 回滚 - 删除失败的绿色版本
        control.delete_agent_runtime(name=f"{agent_name}-green")
        raise Exception("Green deployment failed tests")

Rollback

回滚

bash
undefined
bash
undefined

Quick rollback via CLI

通过CLI快速回滚

agentcore rollback --to-version v1.0.0
agentcore rollback --to-version v1.0.0

Via boto3

通过boto3回滚

control.update_agent_runtime( agentRuntimeId='runtime-xxx', agentRuntimeArtifact={ 's3': {'uri': 's3://bucket/agents/v1.0.0/package.zip'} } )

---
control.update_agent_runtime( agentRuntimeId='runtime-xxx', agentRuntimeArtifact={ 's3': {'uri': 's3://bucket/agents/v1.0.0/package.zip'} } )

---

Related Skills

相关技能

  • bedrock-agentcore: Core platform features
  • bedrock-agentcore-evaluations: Pre-deployment testing
  • terraform-aws: Infrastructure as code
  • ecs-deployment: Alternative deployment patterns
  • bedrock-agentcore：核心平台功能
  • bedrock-agentcore-evaluations：部署前测试
  • terraform-aws：基础设施即代码
  • ecs-deployment：替代部署模式

References

参考资料

  • references/iam-policies.md
    - IAM policy templates
  • references/troubleshooting.md
    - Common deployment issues
  • references/performance-tuning.md
    - Optimization guide
  • references/iam-policies.md
    - IAM策略模板
  • references/troubleshooting.md
    - 常见部署问题
  • references/performance-tuning.md
    - 优化指南

Sources

来源