Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
92 changes: 68 additions & 24 deletions README.md
Original file line number Diff line number Diff line change
@@ -1,14 +1,28 @@
# Etch

Etch is a small notification hub built on Kafka. You place an order, the order service answers right away, and the emails and texts go out in the background through their own services.
Distributed Notification Hub

<img src="/etch.gif" alt="Etch GIF" width="300">

## What's going on here

Etch is a small notification hub built on Kafka. When someone places an order, the order service answers right away. The emails and texts go out later, in the background, through their own services. So a slow or broken email provider never makes an order fail.

- Place an order, and the answer comes back fast
- Notifications get sent separately, on their own time
- If a send fails, Etch retries it
- If it can never work, Etch parks it safely instead of losing it

## Why I built it

I wanted to understand what happens when the slow, unreliable part of a system sits behind a message queue. Sending a notification is a good example. Email and SMS providers time out, go down, and occasionally reject a message outright, and none of that should make an order fail or make a customer wait.
I wanted to understand what happens when the slow, unreliable part of a system sits behind a message queue. Sending a notification is a good example. Email and SMS providers time out, go down, and sometimes reject a message outright. None of that should make an order fail or make a customer wait.

So I built the whole path to see how it holds up: an order comes in, an event goes onto Kafka, and a separate service picks it up and handles delivery. The parts I cared about most were the unglamorous ones. What happens when a send fails? How many times do you retry? Where does a message go when it can never succeed? And what stops a customer from getting the same email twice when Kafka redelivers an event? Etch is my answer to those questions, kept small enough to read in an afternoon.
So I built the whole path to see how it holds up. I cared most about the unglamorous parts. Etch is my answer to these questions, kept small enough to read in an afternoon.

- What happens when a send fails?
- How many times should you retry?
- Where does a message go when it can never succeed?
- What stops a customer from getting the same email twice when Kafka redelivers an event?

## How it fits together

Expand Down Expand Up @@ -43,36 +57,51 @@ flowchart LR
NOTIFY -->|sent / failed / dead-letter| KAFKA
```

1. The client creates an order through the gateway.
2. The order service saves it to MySQL, publishes an `order-created` event, and responds immediately.
3. The notification service consumes the event and creates one notification per requested channel (email, SMS, or both).
4. Each notification is sent through the email or SMS service. If a send fails temporarily, it's retried with exponential backoff.
5. If it fails for good, or runs out of retries, it goes to a dead-letter topic instead of being dropped.
Here's the journey of one order.

- The client creates an order through the gateway
- The order service saves it to MySQL and publishes an `order-created` event
- It responds right away, without waiting for any notifications
- The notification service picks up the event
- It creates one notification for each channel requested (email, SMS, or both)
- Each notification goes out through the email or SMS service
- If a send fails temporarily, it's retried with exponential backoff
- If it fails for good, or runs out of retries, it goes to a dead-letter topic instead of being dropped

## What it does

- **Async by design.** Creating an order never waits on delivery.
- **Retries and a dead-letter queue.** Temporary failures are retried with backoff. Permanent failures and exhausted retries land in the `notification-dlt` topic and get logged.
- **No duplicates.** Redis tracks which events and notifications have already been handled, so a redelivered Kafka message doesn't send twice.
- **Five services, each in its own container.** Gateway, order, notification, email, and SMS. The email and SMS services are mocks that fail on purpose, so you can watch the failure handling work.
- **Tested.** Unit tests for the logic, plus Testcontainers integration tests that run against real MySQL, Kafka, and Redis.
- **Ships to AWS.** A GitHub Actions pipeline tests the code, builds the Docker images, and deploys to ECS Fargate.
- **Async by design.** Creating an order never waits on delivery
- **Retries and a dead-letter queue.** Temporary failures are retried with backoff, and permanent ones land in the `notification-dlt` topic and get logged
- **No duplicates.** Redis tracks which events and notifications were already handled, so a redelivered Kafka message doesn't send twice
- **Five services, each in its own container.** Gateway, order, notification, email, and SMS
- **Mock providers that fail on purpose.** The email and SMS services are stand-ins, so you can watch the failure handling work
- **Easy to follow in the logs.** Requests carry an `X-Correlation-Id` header
- **Tested.** Unit tests for the logic, plus Testcontainers integration tests against real MySQL, Kafka, and Redis
- **Set up for AWS.** A GitHub Actions pipeline tests the code, builds the Docker images, and deploys to ECS Fargate

## Tech stack

Java 25, Spring Boot, Spring Cloud Gateway, Apache Kafka, MySQL, Redis, JUnit and Testcontainers, Docker and Docker Compose, GitHub Actions, and AWS (ECS/Fargate, ECR, RDS, ElastiCache).
- Java 25
- Spring Boot and Spring Cloud Gateway
- Apache Kafka
- MySQL
- Redis
- JUnit and Testcontainers
- Docker and Docker Compose
- GitHub Actions
- AWS (ECS/Fargate, ECR, RDS, ElastiCache)

## Running it

You need Docker with Compose v2. To build or test outside Docker you also need JDK 25 and Maven 3.9 or newer.
You'll need Docker with Compose v2. To build or test outside Docker, you'll also need JDK 25 and Maven 3.9 or newer.

From the repo root:
From the repo root, start everything.

```bash
docker compose up -d --build
```

That starts Kafka, MySQL, Redis, and all five services. The first build takes a few minutes.
That starts Kafka, MySQL, Redis, and all five services. The first build takes a few minutes. Then you can reach each service here.

| Service | URL |
|---|---|
Expand All @@ -84,21 +113,21 @@ That starts Kafka, MySQL, Redis, and all five services. The first build takes a

### Try it

Create an order through the gateway. Three demo users are created on startup.
Three demo users are created on startup. Start by creating an order through the gateway.

```bash
curl -s -X POST http://localhost:8080/orders \
-H 'Content-Type: application/json' \
-d '{"userId":1,"orderNumber":"ORD-1001","total":49.99,"channels":["EMAIL","SMS"]}'
```

Then look at what happened to its notifications:
Then look at what happened to its notifications.

```bash
curl -s http://localhost:8080/notifications/order/1
```

The mock services reject some requests on purpose, so create a few orders and watch the notification service work through retries and dead-lettering:
The mock services reject some requests on purpose. So create a few more orders, then watch the notification service work through retries and dead-lettering.

```bash
docker compose logs -f notification-service
Expand All @@ -110,11 +139,18 @@ docker compose logs -f notification-service
docker compose down
```

Add `-v` to also wipe the MySQL volume and start clean next time.
Add `-v` to also wipe the MySQL volume, so you start clean next time.

## Endpoints

- `POST /orders` creates an order
- `GET /orders` and `GET /orders/{id}` read orders
- `GET /notifications/order/{orderId}` shows the notifications for an order
- The mock providers listen at `POST /email/send` and `POST /sms/send`

## Configuration

Everything has a sensible local default, and `docker-compose.yml` overrides what needs to change inside containers. These are the ones you're most likely to touch:
Everything has a sensible local default. The `docker-compose.yml` file overrides what needs to change inside containers. These are the ones you're most likely to touch.

| Variable | Used by | Default | What it does |
|---|---|---|---|
Expand All @@ -128,6 +164,8 @@ Everything has a sensible local default, and `docker-compose.yml` overrides what

## Testing

Run these from the repo root.

```bash
mvn test # unit tests
mvn verify # unit tests plus the Testcontainers integration tests (needs Docker)
Expand Down Expand Up @@ -163,4 +201,10 @@ infrastructure/

## Deployment

The repo includes ECS task definitions for each service, targeting ECS Fargate with images in ECR, MySQL on RDS, and Redis on ElastiCache. The GitHub Actions pipeline runs the tests, builds the images, and pushes and deploys them when AWS credentials and repository secrets are configured. See [infrastructure/aws/README.md](infrastructure/aws/README.md) for the setup notes.
The repo is set up to deploy to AWS. The infrastructure itself isn't created by code in this repo, so you set that up once by hand.

- There are ECS task definitions for each service, targeting ECS Fargate
- Images live in ECR, MySQL runs on RDS, and Redis runs on ElastiCache
- The GitHub Actions pipeline runs the tests, builds the images, then pushes and deploys them
- The deploy steps only work once AWS credentials and repository secrets are configured
- The setup notes are in [infrastructure/aws/README.md](infrastructure/aws/README.md)
Original file line number Diff line number Diff line change
Expand Up @@ -33,6 +33,9 @@ public boolean claim(String key) {
}

public boolean claim(String key, Duration ttl) {
if (ttl.isZero() || ttl.isNegative()) {
throw new IllegalArgumentException("Idempotency claim ttl must be positive (was " + ttl + ")");
}
Boolean firstClaim = redisTemplate.opsForValue().setIfAbsent(key, "1", ttl);
return Boolean.TRUE.equals(firstClaim);
}
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,9 @@ public class ExponentialBackOff implements BackOff {
private final long baseDelayMs;

public ExponentialBackOff(@Value("${etch.retry.base-delay-ms:500}") long baseDelayMs) {
if (baseDelayMs <= 0) {
throw new IllegalArgumentException("etch.retry.base-delay-ms must be positive (was " + baseDelayMs + ")");
}
this.baseDelayMs = baseDelayMs;
}

Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -49,6 +49,9 @@ public NotificationDispatchService(List<ChannelClient> channelClients,
NotificationMessageRenderer messageRenderer,
BackOff backOff,
@Value("${etch.retry.max-attempts:3}") int maxAttempts) {
if (maxAttempts < 1) {
throw new IllegalArgumentException("etch.retry.max-attempts must be at least 1 (was " + maxAttempts + ")");
}
this.clientsByChannel = channelClients.stream()
.collect(Collectors.toMap(ChannelClient::channel, Function.identity()));
this.notificationRepository = notificationRepository;
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,7 @@
import java.time.Duration;

import static org.assertj.core.api.Assertions.assertThat;
import static org.assertj.core.api.Assertions.assertThatThrownBy;
import static org.mockito.ArgumentMatchers.any;
import static org.mockito.ArgumentMatchers.eq;
import static org.mockito.Mockito.lenient;
Expand Down Expand Up @@ -58,6 +59,14 @@ void claim_usesTheSuppliedTtlWhenGiven() {
verify(valueOperations).setIfAbsent("k2", "1", customTtl);
}

@Test
void claim_rejectsNonPositiveTtl() {
assertThatThrownBy(() -> idempotencyService.claim("k3", Duration.ZERO))
.isInstanceOf(IllegalArgumentException.class);
assertThatThrownBy(() -> idempotencyService.claim("k3", Duration.ofSeconds(-1)))
.isInstanceOf(IllegalArgumentException.class);
}

@Test
void release_deletesTheKey() {
idempotencyService.release("k1");
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,7 @@
import org.junit.jupiter.api.Test;

import static org.assertj.core.api.Assertions.assertThat;
import static org.assertj.core.api.Assertions.assertThatThrownBy;

class ExponentialBackOffTest {

Expand All @@ -15,6 +16,12 @@ void delayDoublesWithEachAttempt() {
assertThat(backOff.delayMsFor(3)).isEqualTo(400L);
}

@Test
void rejectsNonPositiveBaseDelay() {
assertThatThrownBy(() -> new ExponentialBackOff(0L)).isInstanceOf(IllegalArgumentException.class);
assertThatThrownBy(() -> new ExponentialBackOff(-1L)).isInstanceOf(IllegalArgumentException.class);
}

@Test
void neverProducesANegativeExponentForAttemptZeroOrBelow() {
ExponentialBackOff backOff = new ExponentialBackOff(100L);
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,7 @@
import java.util.List;
import java.util.Optional;

import static org.assertj.core.api.Assertions.assertThatThrownBy;
import static org.mockito.ArgumentMatchers.any;
import static org.mockito.ArgumentMatchers.anyInt;
import static org.mockito.ArgumentMatchers.anyString;
Expand Down Expand Up @@ -58,6 +59,14 @@ void setUp() {
messageRenderer, backOff, MAX_ATTEMPTS);
}

@Test
void constructor_rejectsMaxAttemptsBelowOne() {
assertThatThrownBy(() -> new NotificationDispatchService(
List.of(emailClient), notificationRepository, deadLetterService, outcomeRecorder,
messageRenderer, backOff, 0))
.isInstanceOf(IllegalArgumentException.class);
}

private Notification pendingNotification() {
Notification notification = new Notification(1L, NotificationChannel.EMAIL, "a@example.com", "corr-1");
setId(notification, 10L);
Expand Down
Loading