Skip to content
whchoi98Public

Latest commit

Β 

History

368 Commits

Folders and files

NameName
Last commit message
Last commit date
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

AWSops λŒ€μ‹œλ³΄λ“œ v1.8.0

Steampipe, Next.js 14, Amazon Bedrock AgentCore 기반 AWS + Kubernetes 운영 λŒ€μ‹œλ³΄λ“œ

μ‹€μ‹œκ°„ AWS/K8s λ¦¬μ†ŒμŠ€ λͺ¨λ‹ˆν„°λ§, λ„€νŠΈμ›Œν¬ νŠΈλŸ¬λΈ”μŠˆνŒ…, CIS μ»΄ν”ŒλΌμ΄μ–ΈμŠ€ μŠ€μΊ”, AI 기반 뢄석을 단일 λŒ€μ‹œλ³΄λ“œμ—μ„œ μ œκ³΅ν•©λ‹ˆλ‹€.

ν˜„ν™©: 40 νŽ˜μ΄μ§€ Β· 57 라우트 Β· 25 쿼리 파일 Β· 17 API 라우트 Β· 125 MCP 도ꡬ (8 Gateway) Β· 19 μ»΄ν¬λ„ŒνŠΈ


μ•„ν‚€ν…μ²˜

AWSops Architecture

β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚                              Internet                                        β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
                                  β”‚
                                  v
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚  CloudFront (HTTPS)                                                          β”‚
β”‚  β”Œβ”€ Lambda@Edge (us-east-1) ─────────────────────────────────────────────┐  β”‚
β”‚  β”‚  JWT cookie verification -> Cognito Hosted UI redirect or pass throughβ”‚  β”‚
β”‚  β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜  β”‚
β”‚  /awsops*       -> ALB:3000 (Dashboard)                                     β”‚
β”‚  /*             -> ALB:80   (VSCode)                                        β”‚
β”‚  Security: X-Custom-Secret header                                           β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
                                  β”‚
                                  v
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚  ALB (Internet-facing) β€” SG: CloudFront Prefix List (port 80-3000)           β”‚
β”‚  Port 80 -> VSCode (8888)  |  Port 3000 -> Dashboard (3000)                 β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
                                  β”‚
                                  v
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚  EC2 (t4g.2xlarge, Private Subnet) β€” All services on single instance        β”‚
β”‚                                       λͺ¨λ“  μ„œλΉ„μŠ€κ°€ 단일 μΈμŠ€ν„΄μŠ€μ—μ„œ μ‹€ν–‰     β”‚
β”‚                                                                              β”‚
β”‚  β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”  β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”  β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”  β”‚
β”‚  β”‚  Next.js :3000  β”‚  β”‚  Steampipe :9193 β”‚  β”‚  VSCode :8888             β”‚  β”‚
β”‚  β”‚  (40 Pages)     │──│  (Embedded PG)   β”‚  β”‚  (code-server)            β”‚  β”‚
β”‚  β”‚  (17 APIs)      β”‚  β”‚  aws / k8s / trivyβ”‚  β”‚                           β”‚  β”‚
β”‚  β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜  β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜  β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜  β”‚
β”‚  β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”  β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”  β”‚
β”‚  β”‚  Powerpipe      β”‚  β”‚  Docker (λΉŒλ“œ μ „μš©, 싀행은 AgentCore μ„œλΉ„μŠ€)   β”‚  β”‚
β”‚  β”‚  CIS Benchmark  β”‚  β”‚  awsops-agent 이미지 λΉŒλ“œ β†’ ECR ν‘Έμ‹œ          β”‚  β”‚
β”‚  β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜  β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜  β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
                                  β”‚
                                  v
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚  Amazon Bedrock AgentCore (ap-northeast-2)                                   β”‚
β”‚                                                                              β”‚
β”‚  β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”  β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”  β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”  β”‚
β”‚  β”‚  1 Runtime   β”‚  β”‚  8 Gateways     β”‚  β”‚  19 Lambda Targets            β”‚  β”‚
β”‚  β”‚  (Strands)   │──│  125 MCP Tools  │──│  (boto3, read-only)           β”‚  β”‚
β”‚  β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜  β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜  β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜  β”‚
β”‚  β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”                                        β”‚
β”‚  β”‚  Code Interpreter (Python)      β”‚                                        β”‚
β”‚  β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜                                        β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜

κΈ°λŠ₯

λŒ€μ‹œλ³΄λ“œ νŽ˜μ΄μ§€ (40개) / Dashboard Pages (40 pages)

Category Page Path Features / κΈ°λŠ₯
Overview Dashboard /awsops 18 StatsCards, Live Resources, Charts, Warnings
AI Assistant /awsops/ai Claude Sonnet/Opus 4.6, SSE streaming, multi-route
AgentCore /awsops/agentcore Runtime status, 8 Gateways, 125 tools
Bedrock /awsops/bedrock Model usage, token costs, prompt caching, Account vs AWSops
Accounts /awsops/accounts Multi-account management, target account CRUD (admin only)
Compute EC2 /awsops/ec2 Instances + detail panel
Lambda /awsops/lambda Functions, runtimes, memory/timeout
ECS /awsops/ecs Clusters, services, tasks
ECR /awsops/ecr Repositories, images, scan results
EKS Overview /awsops/k8s Clusters, nodes, pod summary
EKS Pods /awsops/k8s/pods Pod list, status, restart counts
EKS Nodes /awsops/k8s/nodes Node list, capacity, conditions
EKS Deployments /awsops/k8s/deployments Deployment list, replicas
EKS Services /awsops/k8s/services Service list, types, endpoints
EKS Explorer /awsops/k8s/explorer K9s-style terminal UI
ECS Container Cost /awsops/container-cost Fargate pricing, Container Insights metrics
EKS Container Cost /awsops/eks-container-cost OpenCost (CPU/Mem/Net/Storage/GPU) + request-based fallback
Network & CDN VPC / Network /awsops/vpc VPCs, Subnets, SGs, Route Tables, TGW, ELB, NAT, IGW + Resource Map
CloudFront /awsops/cloudfront-cdn Distributions, origins, behaviors
WAF /awsops/waf Web ACLs, rules, metrics
Topology /awsops/topology Infra Map + Graph / K8s Map (React Flow)
Storage & DB EBS /awsops/ebs Volumes, Snapshots, encryption, EC2 attachment mapping
S3 /awsops/s3 Buckets, TreeMap, search, IAM analysis
RDS /awsops/rds Instances, SG chaining, metrics
DynamoDB /awsops/dynamodb Tables, capacity, indexes
ElastiCache /awsops/elasticache Clusters, SG, metrics
OpenSearch /awsops/opensearch Domains, encryption, VPC, CloudWatch metrics
MSK /awsops/msk Clusters, broker nodes, CPU/Memory/Network metrics
Datasources Datasources /awsops/datasources 7 platforms CRUD, connection test, auth config / 7μ’… ν”Œλž«νΌ 관리
Explore /awsops/datasources/explore Direct query + AI query generation / 직접 쿼리 + AI 쿼리 생성
Diagnosis AI Diagnosis /awsops/ai-diagnosis 15-section AI analysis, PPTX/DOCX/PDF export / 15μ„Ήμ…˜ AI 진단
Report /awsops/ai-diagnosis/report Print-friendly report (A4, Print-to-PDF) / μΈμ‡„μš© 리포트
Monitoring Monitoring /awsops/monitoring CPU, Memory, Network, Disk I/O (date range)
CloudWatch /awsops/cloudwatch Alarms, state history
CloudTrail /awsops/cloudtrail Trails, events (read/write)
Cost Explorer /awsops/cost Period/service filter, daily/monthly breakdown, MSP auto-detect
Resource Inventory /awsops/inventory Resource count trends, cost impact estimation
Security IAM /awsops/iam Users, roles, trust policies
Security /awsops/security Public S3, Open SGs, Unencrypted EBS, CVE
CIS Compliance /awsops/compliance CIS v1.5~v4.0 benchmarks (431 controls)

AI μ–΄μ‹œμŠ€ν„΄νŠΈ

11단계 라우트 λΆ„λ₯˜ / 11-Route Classification

The AI classifier analyzes user questions and routes them to 1-3 optimal gateways in parallel, then synthesizes the results. AI λΆ„λ₯˜κΈ°κ°€ μ‚¬μš©μž μ§ˆλ¬Έμ„ λΆ„μ„ν•˜μ—¬ 1~3개의 졜적 κ²Œμ΄νŠΈμ›¨μ΄λ‘œ 병렬 λΌμš°νŒ…ν•œ ν›„ κ²°κ³Όλ₯Ό ν†΅ν•©ν•©λ‹ˆλ‹€.

User Question / μ‚¬μš©μž 질문
    |
    |-- "Run code", "calculate"  --> Code --> Bedrock + Code Interpreter (Python sandbox)
    |
    |-- VPC, TGW, VPN, ENI      --> Network --> AgentCore Runtime (17 tools)
    |
    |-- EKS, ECS, Istio         --> Container --> AgentCore Runtime (24 tools)
    |
    |-- CDK, Terraform, CFn     --> IaC --> AgentCore Runtime (12 tools)
    |
    |-- DynamoDB, RDS, Cache    --> Data --> AgentCore Runtime (24 tools)
    |
    |-- IAM, SG, compliance     --> Security --> AgentCore Runtime (14 tools)
    |
    |-- CloudWatch, alarms, logs --> Monitoring --> AgentCore Runtime (16 tools)
    |
    |-- Cost, budget, savings   --> Cost --> AgentCore Runtime (9 tools)
    |
    |-- Prometheus, Loki, Tempo --> Datasource --> External platforms (7 types)
    |
    |-- EC2, S3, Lambda list    --> AWS Data --> Steampipe SQL + Bedrock analysis
    |
    |-- General questions       --> General --> Ops Gateway (9 tools) + Bedrock fallback

λͺ¨λΈ

  • Claude Sonnet 4.6 (global.anthropic.claude-sonnet-4-6) β€” Fast responses / λΉ λ₯Έ 응닡 (default)
  • Claude Opus 4.6 (global.anthropic.claude-opus-4-6-v1) β€” Deep analysis / 심측 뢄석

8 AgentCore Gateways (125 MCP Tools)

Gateway Lambda Targets Tools Key Capabilities / μ£Όμš” κΈ°λŠ₯
Network network-mcp, reachability, flowmonitor 17 VPC, TGW, VPN, ENI, Reachability Analyzer, Flow Logs
Container eks-mcp, ecs-mcp, istio-mcp 24 EKS cluster/node/pod, ECS service/task, Istio mesh
IaC iac-mcp, terraform-mcp 12 CloudFormation validate, CDK docs, Terraform modules
Data dynamodb-mcp, rds-mcp, valkey-mcp, msk-mcp 24 DynamoDB query, RDS Data API, ElastiCache, MSK Kafka
Security iam-mcp 14 IAM users/roles/policies, policy simulation, MFA
Monitoring cloudwatch-mcp, cloudtrail-mcp 16 Metrics, alarms, Log Insights, CloudTrail events
Cost cost-mcp 9 Cost Explorer, Pricing, Budgets, forecasts
Ops knowledge, core-mcp, steampipe-query 9 AWS docs, CLI execution, Steampipe SQL
Total 19 Targets 125

μ£Όμš” AI κΈ°λŠ₯

  • Multi-route: Classifier returns 1-3 routes, parallel gateway calls + synthesis / 1~3개 라우트 병렬 호좜 + κ²°κ³Ό 톡합
  • SSE streaming: Real-time response delivery / μ‹€μ‹œκ°„ 응닡 전달
  • Code Interpreter: Python sandbox for calculations and visualizations / 계산 및 μ‹œκ°ν™”μš© Python μƒŒλ“œλ°•μŠ€
  • Conversation history: Context-aware follow-up questions / λŒ€ν™” νžˆμŠ€ν† λ¦¬ 기반 λ§₯락 μœ μ§€

데이터 흐름

AWSops Flow

β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”     β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”     β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ Browser  β”‚     β”‚ Next.js :3000   β”‚     β”‚ Steampipe (Embedded PG)     β”‚
β”‚          │────>β”‚ POST /awsops/   │────>β”‚ :9193                        β”‚
β”‚ 40 Pages β”‚     β”‚  api/steampipe  β”‚     β”‚                              β”‚
β”‚ Charts   β”‚     β”‚ batchQuery()    β”‚     β”‚ |- aws (380+ tables)  -> AWS API
β”‚ Tables   β”‚<────│ 3 sequential    β”‚<────│ |- k8s (60+ tables)   -> K8s API
β”‚          β”‚     β”‚ 5min TTL cache  β”‚     β”‚ |- trivy              -> CVE DB
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜     β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜     β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
Path / 경둜 Data Source / 데이터 μ†ŒμŠ€ Response Time / 응닡 μ‹œκ°„
Dashboard pages Steampipe pg Pool -> AWS API ~2s (instant with cache)
AI (AWS resources) Steampipe + Bedrock Sonnet 4.6 ~5s
AI (network analysis) AgentCore -> Gateway MCP -> Lambda ~30-60s
AI (code execution) Bedrock + Code Interpreter ~10s
CIS Compliance Powerpipe -> Steampipe -> AWS API ~3-5min
Data Analytics DynamoDB/RDS/ElastiCache/MSK -> Data Gateway ~10-30s
Topology graph Steampipe -> React Flow ~2s

μŠ€ν¬λ¦°μƒ·

Dashboard / λŒ€μ‹œλ³΄λ“œ

Dashboard

AI Assistant / AI μ–΄μ‹œμŠ€ν„΄νŠΈ

AI Assistant

EC2 Instances / EC2 μΈμŠ€ν„΄μŠ€

EC2

EKS Overview / EKS κ°œμš”

EKS

Cost Explorer / λΉ„μš© 뢄석

Cost


λΉ λ₯Έ μ‹œμž‘

사전 μš”κ΅¬μ‚¬ν•­

  • AWS Account (Admin access)
  • EC2 Instance (Amazon Linux 2023, t4g.2xlarge+)
  • AWS credentials configured
  • kubectl + kubeconfig (for K8s features / K8s κΈ°λŠ₯용)

μ„€μΉ˜ / Installation

# Step 0: Deploy CDK infrastructure (run from local machine)
# CDK 인프라 배포 (둜컬 λ¨Έμ‹ μ—μ„œ μ‹€ν–‰)
export VSCODE_PASSWORD='YourPassword'
bash scripts/00-deploy-infra.sh
#   -> cdk bootstrap + cdk deploy AwsopsStack
#   -> VPC, EC2, ALB, CloudFront, SSM Endpoints

# Connect to EC2 via SSM / SSM으둜 EC2 접속
aws ssm start-session --target <instance-id>

# Step 1-3: Install dashboard (inside EC2)
# λŒ€μ‹œλ³΄λ“œ μ„€μΉ˜ (EC2 λ‚΄λΆ€)
cd /home/ec2-user/awsops
bash scripts/install-all.sh        # 01->02->03->11 auto execution

# Step 4: EKS access setup (optional)
# EKS μ ‘κ·Ό μ„€μ • (선택사항)
bash scripts/04-setup-eks-access.sh

# Step 5: Cognito authentication
# Cognito 인증 μ„€μ •
bash scripts/05-setup-cognito.sh

# Step 6: AgentCore AI (batch or individual)
# AgentCore AI (일괄 λ˜λŠ” κ°œλ³„ μ‹€ν–‰)
bash scripts/06-setup-agentcore.sh           # 6a->6b->6c->6d->6e batch
  # Or run individually / λ˜λŠ” κ°œλ³„ μ‹€ν–‰:
  # bash scripts/06a-setup-agentcore-runtime.sh      # Runtime
  # bash scripts/06b-setup-agentcore-gateway.sh      # Gateway
  # bash scripts/06c-setup-agentcore-tools.sh        # Lambda + MCP (19 Lambdas, 8 Gateways)
  # bash scripts/06d-setup-agentcore-interpreter.sh  # Code Interpreter
  # bash scripts/06e-setup-agentcore-config.sh       # Runtime config

# Step 8: Lambda@Edge -> CloudFront integration
# Lambda@Edge -> CloudFront 연동
bash scripts/08-setup-cloudfront-auth.sh

# Step 13: Steampipe systemd unit (Restart=always, boot-time start)
# Steampipe systemd μœ λ‹› 등둝 (μžλ™ μž¬μ‹œμž‘, λΆ€νŒ… μ‹œ μžλ™ μ‹œμž‘)
sudo bash scripts/13-setup-steampipe-systemd.sh

운영

bash scripts/09-start-all.sh    # Start + status + URLs (systemctl-aware)
bash scripts/10-stop-all.sh     # Stop all services (systemctl-aware)
bash scripts/11-verify.sh       # Health check

Step 13 적용 ν›„ SteampipeλŠ” sudo systemctl {start|stop|restart} steampipe둜 κ΄€λ¦¬ν•©λ‹ˆλ‹€ β€” CLI steampipe service stop은 Restart=alwaysκ°€ μžλ™μœΌλ‘œ λ˜λŒλ¦½λ‹ˆλ‹€. After Step 13, manage Steampipe via sudo systemctl {start|stop|restart} steampipe β€” a bare CLI steampipe service stop is auto-undone by Restart=always.


ν”„λ‘œμ νŠΈ ꡬ쑰

awsops/
β”œβ”€β”€ src/
β”‚   β”œβ”€β”€ app/                      # 40 pages + 17 API routes
β”‚   β”‚   β”œβ”€β”€ page.tsx              # Dashboard home (18 StatsCards)
β”‚   β”‚   β”œβ”€β”€ ai/                   # AI Assistant (SSE streaming)
β”‚   β”‚   β”œβ”€β”€ ec2/                  # EC2 instances
β”‚   β”‚   β”œβ”€β”€ lambda/               # Lambda functions
β”‚   β”‚   β”œβ”€β”€ ecs/                  # ECS clusters/services
β”‚   β”‚   β”œβ”€β”€ ecr/                  # ECR repositories
β”‚   β”‚   β”œβ”€β”€ k8s/                  # EKS (overview, pods, nodes, deployments, services, explorer)
β”‚   β”‚   β”œβ”€β”€ vpc/                  # VPC/Network (VPCs, Subnets, SGs, TGW, ELB, NAT, IGW)
β”‚   β”‚   β”œβ”€β”€ cloudfront-cdn/       # CloudFront distributions
β”‚   β”‚   β”œβ”€β”€ waf/                  # WAF Web ACLs
β”‚   β”‚   β”œβ”€β”€ topology/             # Infra Map + K8s Map (React Flow)
β”‚   β”‚   β”œβ”€β”€ ebs/                  # EBS volumes/snapshots (encryption, attachments)
β”‚   β”‚   β”œβ”€β”€ s3/                   # S3 buckets (TreeMap, IAM)
β”‚   β”‚   β”œβ”€β”€ rds/                  # RDS instances (SG chaining)
β”‚   β”‚   β”œβ”€β”€ dynamodb/             # DynamoDB tables
β”‚   β”‚   β”œβ”€β”€ elasticache/          # ElastiCache clusters
β”‚   β”‚   β”œβ”€β”€ monitoring/           # CPU/Memory/Network/Disk (date range)
β”‚   β”‚   β”œβ”€β”€ cloudwatch/           # CloudWatch alarms
β”‚   β”‚   β”œβ”€β”€ cloudtrail/           # CloudTrail events
β”‚   β”‚   β”œβ”€β”€ opensearch/            # OpenSearch 도메인 (domains, encryption, VPC, CW metrics)
β”‚   β”‚   β”œβ”€β”€ msk/                  # MSK Kafka ν΄λŸ¬μŠ€ν„° (brokers, CW metrics: CPU/Memory/Net)
β”‚   β”‚   β”œβ”€β”€ container-cost/       # ECS Container Cost (Fargate pricing)
β”‚   β”‚   β”œβ”€β”€ eks-container-cost/  # EKS Container Cost (OpenCost + request-based)
β”‚   β”‚   β”œβ”€β”€ cost/                 # Cost Explorer (period/service, MSP auto-detect)
β”‚   β”‚   β”œβ”€β”€ inventory/            # Resource Inventory (count trends, cost impact)
β”‚   β”‚   β”œβ”€β”€ iam/                  # IAM users/roles
β”‚   β”‚   β”œβ”€β”€ security/             # Security findings
β”‚   β”‚   β”œβ”€β”€ compliance/           # CIS v1.5~v4.0 benchmarks
β”‚   β”‚   β”œβ”€β”€ datasources/          # External datasources (7 platforms CRUD + Explore)
β”‚   β”‚   β”œβ”€β”€ ai-diagnosis/         # AI comprehensive diagnosis (15 sections + print report)
β”‚   β”‚   └── api/                  # 17 API routes (ai, steampipe, auth, msk, rds, elasticache, opensearch, agentcore, code, benchmark, container-cost, eks-container-cost, bedrock-metrics, datasources, k8s, report, notification)
β”‚   β”œβ”€β”€ components/               # 19 shared components (Sidebar, Charts, Table, K8s, AccountSelector, ReportMarkdown)
β”‚   β”œβ”€β”€ contexts/                # AccountContext (multi-account state)
β”‚   β”œβ”€β”€ lib/steampipe.ts          # pg Pool (NOT CLI) β€” max 10, 30s statement + 40s client timeout, 5min cache
β”‚   β”œβ”€β”€ lib/resource-inventory.ts  # λ¦¬μ†ŒμŠ€ 인벀토리 μŠ€λƒ…μƒ· (resource snapshots)
β”‚   β”œβ”€β”€ lib/cost-snapshot.ts      # Cost 데이터 μŠ€λƒ…μƒ· (cost data fallback)
β”‚   β”œβ”€β”€ lib/app-config.ts         # μ•± μ„€μ • (app config: costEnabled)
β”‚   β”œβ”€β”€ lib/queries/              # 25 SQL query files (ec2, ebs, msk, opensearch, vpc, s3, rds, k8s, container-cost, eks-container-cost, bedrock...)
β”‚   └── types/aws.ts              # TypeScript type definitions
β”œβ”€β”€ agent/                        # Strands Agent μ†ŒμŠ€ (EC2μ—μ„œ λΉŒλ“œ β†’ ECR β†’ AgentCoreμ—μ„œ μ‹€ν–‰)
β”‚   β”œβ”€β”€ agent.py                  # Main entrypoint with dynamic gateway selection
β”‚   β”œβ”€β”€ streamable_http_sigv4.py  # MCP StreamableHTTP with SigV4
β”‚   β”œβ”€β”€ Dockerfile                # Python 3.11-slim, arm64 (EC2μ—μ„œ λΉŒλ“œ, AgentCore Runtimeμ—μ„œ μ‹€ν–‰)
β”‚   └── lambda/                   # 19 Lambda source files + create_targets.py
β”œβ”€β”€ powerpipe/                    # CIS Benchmark mod
β”œβ”€β”€ infra-cdk/                    # CDK TypeScript (AwsopsStack, CognitoStack)
β”‚   └── lib/
β”‚       β”œβ”€β”€ awsops-stack.ts       # VPC, EC2, ALB, CloudFront
β”‚       └── cognito-stack.ts      # Cognito User Pool, Lambda@Edge
β”œβ”€β”€ scripts/                      # 17 install/ops scripts
β”‚   β”œβ”€β”€ 00-deploy-infra.sh        # Step 0: CDK infrastructure
β”‚   β”œβ”€β”€ 01-install-base.sh        # Step 1: Steampipe + Powerpipe
β”‚   β”œβ”€β”€ 02-setup-nextjs.sh        # Step 2: Next.js setup
β”‚   β”œβ”€β”€ 03-build-deploy.sh        # Step 3: Production build
β”‚   β”œβ”€β”€ 04-setup-eks-access.sh    # Step 4: EKS access entry
β”‚   β”œβ”€β”€ 05-setup-cognito.sh       # Step 5: Cognito auth
β”‚   β”œβ”€β”€ 06-setup-agentcore.sh     # Step 6: Wrapper (6a->6b->6c->6d->6e)
β”‚   β”œβ”€β”€ 06a~06e-setup-agentcore-* # Step 6a-6e: AgentCore (split)
β”‚   β”œβ”€β”€ 06f-setup-agentcore-memory.sh  # Step 6f: Memory Store (365-day retention)
β”‚   β”œβ”€β”€ 07-setup-opencost.sh            # Step 7: Prometheus + OpenCost (EKS cost)
β”‚   β”œβ”€β”€ 08-setup-cloudfront-auth.sh # Step 8: Lambda@Edge
β”‚   β”œβ”€β”€ 09-start-all.sh           # Start all services (systemctl-aware)
β”‚   β”œβ”€β”€ 10-stop-all.sh            # Stop all services (systemctl-aware)
β”‚   β”œβ”€β”€ 11-verify.sh              # Health check
β”‚   β”œβ”€β”€ 12-setup-multi-account.sh  # Step 12: Multi-account setup
β”‚   β”œβ”€β”€ 13-setup-steampipe-systemd.sh  # Step 13: Steampipe systemd unit (Restart=always)
β”‚   β”œβ”€β”€ install-all.sh            # Auto: 01->02->03->10
β”‚   └── ARCHITECTURE.md           # Full architecture documentation
β”œβ”€β”€ docs/                         # Guides + Troubleshooting
β”‚   β”œβ”€β”€ INSTALL_GUIDE.md
β”‚   β”œβ”€β”€ TROUBLESHOOTING.md
β”‚   └── decisions/                # Architecture Decision Records (ADR)
β”œβ”€β”€ .kiro/rules.md                # Kiro vibe-coding rules
└── .amazonq/rules.md             # Amazon Q rules

기술 μŠ€νƒ

Layer Technology
Frontend Next.js 14 (App Router), TypeScript, Tailwind CSS (dark navy theme), Recharts, React Flow
Backend Node.js 20, pg (PostgreSQL client), node-cache
Data Steampipe (embedded PostgreSQL, 380+ AWS tables, 60+ K8s tables), Powerpipe
AI Amazon Bedrock (Claude Sonnet/Opus 4.6, ap-northeast-2 global.*), AgentCore Runtime (Strands), 8 AgentCore Gateways (MCP), Code Interpreter
Auth Amazon Cognito (User Pool + Hosted UI), Lambda@Edge (JWT, Python 3.12)
IaC CDK TypeScript (infra-cdk/) β€” AwsopsStack, CognitoStack
Container Docker (arm64), ECR
Serverless 19 Lambda functions (Python 3.12, boto3)

μ‚¬μš©λœ AWS μ„œλΉ„μŠ€

Service Region Purpose / μš©λ„
EC2 (t4g.2xlarge) ap-northeast-2 All services hosting / 전체 μ„œλΉ„μŠ€ ν˜ΈμŠ€νŒ…
ALB ap-northeast-2 Load balancer / λ‘œλ“œλ°ΈλŸ°μ„œ
CloudFront Global CDN + HTTPS
Cognito ap-northeast-2 User authentication / μ‚¬μš©μž 인증
Lambda@Edge us-east-1 CloudFront auth / CloudFront 인증
Lambda (x19) ap-northeast-2 MCP tools (Network, Container, IaC, Data, Security, Monitoring, Cost, Ops)
AgentCore Runtime ap-northeast-2 Strands AI Agent
AgentCore Gateway (x8) ap-northeast-2 MCP tool routing / MCP 도ꡬ λΌμš°νŒ…
AgentCore Code Interpreter ap-northeast-2 Python Sandbox
Bedrock (Sonnet/Opus 4.6) ap-northeast-2 AI models (global.* cross-region inference)
ECR ap-northeast-2 Agent Docker image
DynamoDB ap-northeast-2 Table data queries
ElastiCache ap-northeast-2 Redis/Memcached cluster monitoring
MSK ap-northeast-2 Kafka cluster monitoring
SSM ap-northeast-2 EC2 access
CDK (CloudFormation) ap-northeast-2 Infrastructure deployment (AwsopsStack)
OpenCost + Prometheus EKS cluster Pod-level cost analysis (CPU/Mem/Net/Storage/GPU)

인증

Method Component
CloudFront -> Lambda@Edge JWT cookie (1h TTL)
Cognito Hosted UI OAuth2 Authorization Code
ALB X-Custom-Secret header
AgentCore Gateway IAM Role

μ•Œλ €μ§„ 이슈 및 해결법

See docs/TROUBLESHOOTING.md for details. μžμ„Έν•œ λ‚΄μš©μ€ docs/TROUBLESHOOTING.md μ°Έμ‘°.

Issue / 이슈 Solution / 해결법
SCP blocks IAM/Lambda hydrate ignore_error_codes + remove affected columns
Steampipe CLI slow (4s/query) Use pg Pool (0.006s/query)
basePath not applied to fetch Add /awsops prefix to all fetch URLs
CloudTrail >60s timeout Event tab lazy-load
AgentCore arm64 only docker buildx --platform linux/arm64
PostgreSQL separate install? Not needed -- embedded in Steampipe
CloudFront CachePolicy TTL=0 + Header Use managed CACHING_DISABLED
ALB SG rule limit (CF prefix 120+) Single rule with port range 80-3000
Gateway Target CLI inlinePayload error Use Python/boto3 (mcp.lambda structure)
Code Interpreter name hyphens Underscores only ([a-zA-Z][a-zA-Z0-9_])
Istio resource queries Use Steampipe K8s CRD tables (kubernetes_custom_resource)
VPC Lambda Steampipe access Place Lambda in VPC + allow SG inbound

λ¬Έμ„œ


AWSops Dashboard v1.8.0 (English)

AWS + Kubernetes Operations Dashboard β€” Steampipe, Next.js 14, Amazon Bedrock AgentCore

Real-time AWS/K8s resource monitoring, network troubleshooting, CIS compliance scanning, and AI-powered analysis in a single dashboard.

Stats: 40 pages Β· 57 routes Β· 25 query files Β· 17 API routes Β· 125 MCP tools (8 Gateways) Β· 19 components

Documentation


License

Apache-2.0

Releases

Packages

Contributors

Languages