AGENTS.md

Quick Reference for Vector Development

This guide provides quick commands and coding conventions for Vector development. It's designed to help both AI assistants and human contributors get started quickly.

For comprehensive information, see CONTRIBUTING.md and docs/DEVELOPING.md.

Project Summary

Vector is a high-performance, end-to-end observability data pipeline written in Rust. It collects, transforms, and routes logs, metrics, and traces from various sources to any destination. Vector is designed to be reliable, fast, and vendor-neutral, enabling dramatic cost reduction and improved data quality for observability infrastructure.

Project Structure

Core Directories

• /src/ - Main Rust source code

  • sources/ - Data ingestion components
  • transforms/ - Data processing and routing components
  • sinks/ - Data output destinations
  • config/ - Configuration system and validation
  • topology/ - Component graph management
  • api/ - GraphQL API for management and monitoring
  • cli.rs - Command-line interface

• /lib/ - Modular library crates

  • vector-lib/ - Unified library re-exporting core Vector components
  • vector-core/ - Core event system and abstractions
  • vector-config/ - Configuration framework with schema generation
  • vector-buffers/ - Buffering and backpressure management
  • codecs/ - Data encoding/decoding (JSON, Avro, Protobuf)
  • enrichment/ - Data enrichment (GeoIP, custom tables)
  • file-source/ - File watching and reading
  • prometheus-parser/ - Prometheus metrics parsing

• /config/ - Configuration examples and templates

• /distribution/ - Packaging and deployment configs

  • docker/ - Docker images (Alpine, Debian, Distroless)
  • kubernetes/ - Kubernetes manifests
  • systemd/ - SystemD service files
  • debian/, rpm/ - Linux package configurations

• /scripts/ - Build, test, and deployment automation

• /docs/ - Developer documentation

• /tests/ - Integration and E2E tests

Development Workflow

Iterative Development Process

When working on Vector's Rust codebase, follow this iterative development cycle:

1. Make code changes
2. Run make check-clippy to check for linting issues
3. Fix any issues found (use make clippy-fix for auto-fixes)
4. Continue to next task or mark current task complete

Run this cycle after any code modification.

When editing markdown files (*.md), run make check-markdown after changes.

Two Different Workflows

Rust Development (Most Common)

If you're working on Vector's Rust codebase (sources, sinks, transforms, core functionality):

Format your code:

make fmt

Check formatting:

make check-fmt

Run Clippy (linter):

make check-clippy

Auto-fix Clippy issues:

make clippy-fix

Run unit tests:

make test
# or
cargo nextest run --workspace --no-default-features --features "${FEATURES}"

Run integration tests:

# See available integration tests:
# cargo vdev int show
./scripts/run-integration-test.sh <integration-name>

See Integration Tests section below for more details.

Before committing (recommended checks):

make fmt                      # Format code
make check-fmt                # Verify formatting
make check-clippy             # Run Clippy linter
make check-markdown           # Check markdown files
make check-component-docs     # Check component documentation
./scripts/check_changelog_fragments.sh  # Verify changelog

Website/Docs Development (Separate Process)

If you're working on vector.dev website or documentation content:

Prerequisites:

• Hugo static site generator
• CUE CLI tool
• Node.js and Yarn
• htmltest

Run the site locally:

cd website
make serve
# Navigate to http://localhost:1313

Build website:

cd website
make cue-build

Note: Website changes use Hugo, CUE, Tailwind CSS, and TypeScript. See website/README.md for details.

Common Patterns

Development Tools

Vector uses cargo vdev for most development tasks. This is a custom CLI tool that wraps common operations:

cargo vdev check rust         # Clippy
cargo vdev check fmt          # Formatting check
cargo vdev check events       # Event instrumentation check
cargo vdev check licenses     # License compliance
cargo vdev test               # Unit tests
cargo vdev int test <name>    # Integration tests
cargo vdev fmt                # Format code

Pre-Push Hook (Optional but Recommended)

Create .git/hooks/pre-push with:

#!/bin/sh
set -e

echo "Format code"
make fmt

echo "Running pre-push checks..."
make check-licenses
make check-fmt
make check-clippy
make check-markdown
make check-component-docs

./scripts/check_changelog_fragments.sh

Then: chmod +x .git/hooks/pre-push

Detailed Documentation

Topic	Document
Rust style patterns	docs/RUST_STYLE.md

Architecture Notes

Component Development

• Sources: Ingest data from external systems
• Transforms: Modify, filter, or enrich event data
• Sinks: Send data to external systems

Component docs are auto-generated from code annotations. Run make check-component-docs after changes.

Integration Tests

Integration tests verify Vector works with real external services. Require Docker or Podman.

Run integration tests:

# List available tests
cargo vdev int show

# Run specific test (example: aws)
cargo vdev int start aws # need to initiate dev environment first
cargo vdev int test aws

See docs/DEVELOPING.md for adding new integration tests.

Key Files

• Makefile - Common build/test/check targets
• vdev/ - Custom development CLI tool
• src/ - Rust source code
• website/ - Hugo-based documentation site
• tests/ - Integration and behavior tests

Common Issues

Formatting Fails

Run make fmt before committing. Formatting must be exact.

Clippy Errors

Run make clippy-fix to auto-fix many issues. Manual fixes may be required.

Component Docs Out of Sync

Component documentation is generated from code. Run:

make check-component-docs

License Check Fails

After adding/updating dependencies:

cargo install dd-rust-license-tool --locked
make build-licenses

Reference Documentation

These documents provide context that AI agents and developers need when working on Vector code.

Essential for Code Changes

• STYLE.md - Code style rules (formatting, const strings, code organization)
• docs/ARCHITECTURE.md - System architecture (sources, transforms, sinks, topology)
• docs/DEVELOPING.md - Development workflow and testing

Component Development

• docs/specs/component.md - Component specification (naming, configuration, health checks)
• docs/specs/instrumentation.md - Instrumentation requirements (event/metric naming)
• src/internal_events - Internal event examples for telemetry

Adding Documentation

• docs/DOCUMENTING.md - How to document code changes
• changelog.d/README.md - Adding changelog entries

Full Guides

• CONTRIBUTING.md - Complete contributing guide
• website/README.md - Website development only (separate from Rust code)