AWS DevOps Agent Ops Intelligence Platform

A centralized enterprise platform that aggregates AWS DevOps Agent investigations across multiple AWS accounts, enriches them with business context, and gives engineering leadership real-time operational visibility.

Architecture

Member Account A ──► EventBridge ──┐
Member Account B ──► EventBridge ──┼──► Central EventBridge Bus
Member Account N ──► EventBridge ──┘            │
                                                 ▼
                                        Lambda Aggregator
                                        (enrich + store)
                                                 │
                                    ┌────────────┼────────────┐
                                    ▼            ▼            ▼
                               DynamoDB       SNS Topic    SNS Topic
                            (incident store)     │              │
                                                 ▼              ▼
                                           Slack Bot      Runbook Gen
                                         (team alerts)  (→ Confluence)

Components

Path	Description
aggregator/lambda_handler.py	Multi-account EventBridge ingestion + business context enrichment
bot/slack_handler.py	Slack bot — routes P0/P1 to team + central ops channels
runbook-gen/generator.py	Auto-generates Confluence runbooks from root cause findings
infra/app.py	AWS CDK stack — full IaC deployment

Quick Start

# 1. Clone and install
git clone https://github.com/yourorg/devops-agent-platform
cd devops-agent-platform
pip install aws-cdk-lib constructs

# 2. Store secrets in SSM Parameter Store
aws ssm put-parameter --name /devops-agent/slack-token       --value "xoxb-..." --type SecureString
aws ssm put-parameter --name /devops-agent/confluence-url    --value "https://yourorg.atlassian.net" --type String
aws ssm put-parameter --name /devops-agent/confluence-token  --value "your-token" --type SecureString

# 3. Deploy
cd infra
cdk bootstrap --context centralAccount=123456789012 --context region=us-east-1
cdk deploy    --context centralAccount=123456789012 --context region=us-east-1

# 4. Configure member accounts to forward DevOps Agent events
# In each member account, create an EventBridge rule targeting the central bus ARN

Environment Variables

Variable	Description
DYNAMODB_TABLE	DynamoDB table name for incident storage
ENRICHED_TOPIC_ARN	SNS topic ARN for fan-out
SLACK_BOT_TOKEN	Slack bot OAuth token
SLACK_CHANNEL_ID	Default ops channel (e.g. #ops-incidents)
CONFLUENCE_BASE_URL	Confluence base URL
CONFLUENCE_API_TOKEN	Confluence API token
CONFLUENCE_SPACE_KEY	Space key for runbooks (default: RUNBOOKS)
CONFLUENCE_PARENT_PAGE_ID	Optional parent page ID

Service Catalog

Edit SERVICE_CATALOG in aggregator/lambda_handler.py to map your resource naming conventions to business services, criticality tiers, SLA thresholds, and team ownership.

SERVICE_CATALOG = {
    "payment": {"name": "Payment Service", "criticality": "P0", "sla_minutes": 15, "team": "payments-eng"},
    # Add your services here...
}

Cost Model

DevOps Agent is billed at $0.0083/agent-second. This platform tracks per-investigation cost in DynamoDB and surfaces it in Slack alerts, giving ops teams full cost visibility.

License

MIT