How awslabs/aws-config-resource-schema works
Project type & purpose (what you’re diagramming)
This repo is not an application (no runtime services, no “backend/frontend”). It’s a data/schema artifact repository: a curated set of JSON “resource property” schema files for AWS Config Configuration Items (CIs) that are queryable via the AWS Config SelectResources API.
• Static artifacts (the JSON property/type maps)
• How consumers use them to build valid SelectResources SQL-like queries
• (Optionally) How maintainers update/publish the artifacts (GitHub workflow, review, release), even if the CI details aren’t explicit in the file tree.
From the file tree + README, the core structure is:
• config/properties/resource-types/*.properties.json
One file per AWS Config resource type, named like AWS::EC2::Instance.properties.json.
Each file is effectively a dictionary: property.path → type (string/boolean/integer/etc.).
• config/properties/AWS.properties.json
A merged/aggregated property file containing properties for all supported resource types.
• .github/*, CONTRIBUTING.md, CODE_OF_CONDUCT.md, LICENSE, NOTICE
This means the “architecture” is fundamentally a schema distribution + consumption flow.
What the README implies about behavior & interactions
The README defines the consumer workflow:
Craft a SelectResources query using those property names/types
Also:
• There’s a supported type system: boolean, cidr_block, date, float, integer, ip, string
• Consumers interact with an external managed service: AWS Config (specifically SelectResources API)
Consumer pulls/reads schema
• User/Tool → GitHub Repo (clone/download)
• User/Tool → resource-types/*.properties.json (lookup properties)
• (Alternative path) User/Tool → AWS.properties.json (global lookup)
Consumer authors query
• Schemas → Query Authoring (property names + types inform query correctness)
Query execution
• User/Tool → AWS Config SelectResources API (submit query)
• AWS Config → Query Results (resourceIds / counts)
Maintenance loop (if diagramming lifecycle)
• AWS resource type changes / new CI fields → Contributors/Maintainers
• Maintainers → PR to GitHub → Schema artifacts updated
Architectural patterns / principles to highlight
This repo follows a “schema-as-data” / reference dataset pattern:
• Version-controlled canonical artifacts
• Per-entity modularization (one file per resource type)
• Aggregation (single merged file for convenience)
• Consumer-driven usage (schemas enable correct API usage; repo itself doesn’t run)
Call out that there is no runtime coupling inside the repo—only data organization and consumer workflows.
Technologies / formats to label explicitly
• JSON (*.properties.json)
• AWS Config SelectResources API (external integration)
• “SQL-like query” (as shown in README examples)
• Type system: boolean/cidr_block/date/float/integer/ip/string
What NOT to overcomplicate
• Don’t invent runtime components (servers, queues, DBs).