Barrage (WIP)

Barrage is lightweight, high-performance testing tool designed to push various types of data to target API server, Kafka broker, or more.

It is built with Rust(on top of k8s), allowing you to simulate high-frequency data traffic or schedule periodic data synchronization.

Key Features (TODO)

• Multi-Protocol Support: Push data via HTTP GET/POST or Kafka topics.
• Structured Configuration: Separate deployment specs (dep.yaml) from traffic logic (traffic.yaml).
• K8s Native: Automatically generates professional Kubernetes manifests with resource limits and independent scaling.

---

Configuration

Configurations are stored in the config/ directory.

1. config/dep.yaml (Deployment Spec)

Defines how the workers should behave in Kubernetes.

instance: 2    # Number of replicas per task
cpu: 100m      # CPU limit/request
mem: 128Mi     # Memory limit/request

2. config/traffic.yaml (Traffic Spec)

Defines the actual data tasks to be performed.

# For HTTP API traffic
- type: http
  host: https://jsonplaceholder.typicode.com/
  path: posts/1
  frequency: 5    # 5 requests per minute
  duration: 1m    # Stop after 1 minute

# For Kafka producing traffic
- type: kafka
  host: http://localhost:9092
  topic: sample_topic
  frequency: 10
  duration: 1h    # Stop after 1 hour

---

Getting Started

1. Manual Local Build

To build the project locally, you need cmake, pkg-config, and openssl installed on your system (for rdkafka support).

# On macOS
brew install cmake pkg-config openssl

cargo build --release

2. Docker Image Preparation

For Kubernetes deployment, build the container image:

docker build -t barrage:latest -f docker/Dockerfile .

---

Command Guide

Barrage provides a simple CLI to manage the entire lifecycle:

Kubernetes Lifecycle

Command	Description
cargo run -- init	Generates K8s manifests in k8s/generated/ based on configs.
cargo run -- serve	Applies the generated manifests to your active K8s context.
cargo run -- stop	Stops the running pods and cleans up K8s resources.

Local Testing

Command	Description
cargo run -- worker	Runs a specific task directly on your local machine.
cargo run -- server	Starts a manual trigger server (Axum) on port 3000.

---

Kubernetes Workflow Example

# 1. Generate manifests (timestamped to force pod restarts)
$ cargo run -- init

# 2. Deploy to cluster (Docker Desktop / MiniKube)
$ cargo run -- serve

# 3. Check logs
$ kubectl logs -l app=barrage -f

# 4. Clean up
$ cargo run -- stop

---

Advanced: Image Updates

If you change the source code and want to see it reflected in K8s:

1. Rebuild the image: docker build -t barrage:latest -f docker/Dockerfile .
2. Run cargo run -- init (this updates the restartedAt timestamp to force a rollout).
3. Run cargo run -- serve.