阅读中文版本

Build a Webhook Delivery with Retry

Webhook Delivery with Retry — a complete walkthrough of webhook delivery, covering delivery queue, backoff scheduler, attempt runner, and testing.

webhook_delivery_retry/
├── src/index.ts
├── src/delivery-queue.ts
├── src/backoff-scheduler.ts
├── tests/webhook_delivery_retry.test.ts
└── package.json

// Delivery queue: webhook delivery core
export function deliveryqueue(input: DeliveryqueueInput): DeliveryqueueResult {
  // Build the delivery queue that anchors webhook delivery and defines the core data shape.
  return processDeliveryqueue(input);
}

// Backoff scheduler — handles webhook delivery Add the backoff scheduler so
export class Backoffscheduler {
  constructor(private opts: BackoffschedulerOptions) {}
  async run(ctx: Context): Promise<BackoffschedulerOutput> {
    return this.execute(ctx);
  }
}

// Attempt runner: webhook delivery stage
export const attemptrunner = (next: Next) => async (req: Request) => {
  const out = await transform(req);
  return next(out);
};

// Result classifier for webhook delivery
export async function resultclassifier(stream: AsyncIterable<Chunk>): Promise<Aggregate> {
  let acc = init();
  for await (const c of stream) acc = merge(acc, c);
  return acc;
}

// Dead letter store: webhook delivery finalization
export function deadletterstore(state: State): Result {
  validate(state);
  return finalize(state);
}

Project overview

This project builds a webhook delivery system whose reason for existing is delivering webhooks with exponential backoff and a dead letter sink so a failing endpoint does not block the queue. The temptation in webhook delivery is to bolt the interesting part onto a working prototype and discover the hard cases in production; this build inverts that order and treats an endpoint that returns 200 to a malformed payload and silently swallows webhooks that never take effect as the first thing designed, not the last.

You will implement Delivery queue, Backoff scheduler, Attempt runner, Result classifier, and Dead letter store, each as a module with one responsibility and a test that proves it in isolation. The webhook delivery pipeline is then the wiring that composes them, and the integration test is where the composition is shown to survive an endpoint that returns 200 to a malformed payload and silently swallows webhooks that never take effect.

The walkthrough is written for a developer who has used a webhook delivery system but has not had to make one honest. It favors explicit data shapes over clever abstractions, and it names the failure each design choice prevents rather than presenting the choice as a style preference.

Each phase ends with something observable — a delivery queue that returns the right shape, a backoff scheduler that rejects bad input, an integration test that goes green — so you can stop at any point and still have a webhook delivery system that works for the cases covered so far.

What you will build

The finished webhook delivery system exposes five capabilities, each with a focused test:

• Delivery queue
• Backoff scheduler
• Attempt runner
• Result classifier
• Dead letter store

Delivery queue is the foundation that every later stage assumes; Backoff scheduler is where webhook delivery correctness lives; Attempt runner handles an endpoint that returns 200 to a malformed payload and silently swallows webhooks that never take effect; Result classifier makes the system observable; and Dead letter store is the graceful-degradation layer that keeps overload from becoming corruption.

The build keeps the webhook delivery logic in ordinary modules you can read in one sitting, and pulls in dependencies only for the parts that are genuinely tedious to hand-roll. That keeps the diff between the tutorial and a real codebase small, and keeps the failure modes visible in code rather than hidden behind a framework.

Architecture and data model

The system is organized around Delivery queue, Backoff scheduler, and Attempt runner, with Result classifier tying them together. Data flows as a queue of deliveries each with an attempt count and a next-try time, moved to a dead letter store after exhausting retries.

Each component is a module with a pure interface where practical: the same input produces the same output, and side effects are pushed to the edges. That discipline is what makes the phases independently testable, because a test for Attempt runner does not need to stand up Delivery queue and Backoff scheduler to exercise the interesting branch.

The data model is deliberately explicit. A queue of deliveries each with an attempt count and a next-try time, moved to a dead letter store after exhausting retries Keeping the shape written down — in a schema, a type, or a fixture — means a change to the shape is a visible diff rather than a silent runtime shift, and it means the integration test can assert against a concrete shape instead of guessing.

Configuration and code are kept apart: the configuration says what webhook delivery should do, the modules say how. A reviewer can change a threshold or a policy without reading the implementation, which keeps the risky changes small and lets the diff stand as the review.

The failure path through Attempt runner is designed before the happy path is finished, because an endpoint that returns 200 to a malformed payload and silently swallows webhooks that never take effect is the case that distinguishes a webhook delivery system from a demo. Every component declares what it does when its inputs are wrong, and the integration test asserts each of those declarations.

Prerequisites and dependencies

Have these ready before you install anything, so the webhook delivery setup is not blocked by a missing tool:

• Node.js 20 LTS or later
• Familiarity with webhook delivery fundamentals
• A local environment to run the test suite

The runtime dependencies are chosen to keep the webhook delivery datapath legible:

• fastify (^5.0.0) — Webhook API
• undici (^6.19.0) — Endpoint dispatch
• redis (^4.7.0) — Queue and dead letters

# setup for webhook-delivery-retry
pnpm init
pnpm add fastify undici redis
pnpm add -D typescript tsx vitest

Versions are pinned to a major range so the webhook delivery install is reproducible; commit the lockfile. The delivery queue reads its operational parameters from the environment and validates them at startup, so a missing value fails the boot rather than the first webhook delivery request.

Directory structure

Keep each webhook delivery component in its own file so it can be tested without starting the others:

webhook_delivery_retry/
├── src/index.ts
├── src/delivery-queue.ts
├── src/backoff-scheduler.ts
├── tests/webhook_delivery_retry.test.ts
└── package.json

Phase 1: Delivery queue

Build the delivery queue that anchors webhook delivery and defines the core data shape.

// Delivery queue: webhook delivery core
export function deliveryqueue(input: DeliveryqueueInput): DeliveryqueueResult {
  // Build the delivery queue that anchors webhook delivery and defines the core data shape.
  return processDeliveryqueue(input);
}

After this phase the delivery queue works in isolation and is covered by a unit test. The delivery queue defines the data shape that every later stage assumes, so getting it right first means the later phases compile against a stable contract.

Treat the delivery queue as the single source of truth for the webhook delivery data shape. Validate inputs at the boundary and reject early, so a malformed value never reaches Backoff scheduler or Attempt runner where its effect would be harder to localize.

Phase 2: Backoff scheduler

Add the backoff scheduler so webhook delivery can delivering webhooks with exponential correctly.

// Backoff scheduler — handles webhook delivery Add the backoff scheduler so
export class Backoffscheduler {
  constructor(private opts: BackoffschedulerOptions) {}
  async run(ctx: Context): Promise<BackoffschedulerOutput> {
    return this.execute(ctx);
  }
}

After this phase the backoff scheduler works in isolation and is covered by a unit test. This is where webhook delivery correctness lives: the backoff scheduler is what makes delivering webhooks with exponential backoff and a dead letter sink so a failing endpoint does not block the queue true, and its test is the proof.

Keep the backoff scheduler pure where you can. A function that takes an input and returns an output is trivial to test and impossible to mock wrong; a class with hidden state is the opposite. Push side effects to the caller and keep the backoff scheduler about the transformation.

Phase 3: Attempt runner

Wire the attempt runner to handle an endpoint that returns without losing data.

// Attempt runner: webhook delivery stage
export const attemptrunner = (next: Next) => async (req: Request) => {
  const out = await transform(req);
  return next(out);
};

After this phase the attempt runner works in isolation and is covered by a unit test. The attempt runner is the difference between a webhook delivery system that works on the happy path and one that survives an endpoint that returns 200 to a malformed payload and silently swallows webhooks that never take effect.

Design the failure before the success. Decide what attempt runner does when an endpoint that returns 200 to a malformed payload and silently swallows webhooks that never take effect: reject, retry, shed, or degrade — and write that test first. A webhook delivery system whose failure behavior is unspecified will pick one at random under pressure, and it is usually the wrong one.

Phase 4: Result classifier

Integrate the result classifier to make webhook delivery observable and testable end to end.

// Result classifier for webhook delivery
export async function resultclassifier(stream: AsyncIterable<Chunk>): Promise<Aggregate> {
  let acc = init();
  for await (const c of stream) acc = merge(acc, c);
  return acc;
}

After this phase the result classifier works in isolation and is covered by a unit test. Observability is not a finishing touch; it is how you know the webhook delivery system is doing what it claims in production.

Emit one structured event per meaningful action, with enough context to localize a failure without a second round-trip. The result classifier is what turns “something is slow” into “the backoff scheduler is slow for inputs matching this shape,” which is the difference between an alert and an investigation.

Phase 5: Dead letter store

Finalize the dead letter store so the webhook delivery system degrades gracefully under overload.

// Dead letter store: webhook delivery finalization
export function deadletterstore(state: State): Result {
  validate(state);
  return finalize(state);
}

After this phase the dead letter store works in isolation and is covered by a unit test. The dead letter store is what keeps overload from becoming corruption: when the webhook delivery system is asked for more than it can do, it degrades in a defined way instead of failing in an undefined one.

Pick a degradation strategy for dead letter store — shed, queue, or backpressure — and make it explicit in code and in a test. Leaving the behavior to chance is how a webhook delivery traffic spike becomes a data loss incident.

Testing

Each phase has a focused test: delivery queue, backoff scheduler, and attempt runner are exercised in isolation, then an integration test drives the full webhook delivery pipeline and asserts the an endpoint that returns case is handled.

Start with unit tests that pin each component’s contract: the delivery queue returns the documented shape, the backoff scheduler transforms correctly, and the attempt runner fails in the documented way. These tests are fast and they localize regressions to the component that broke.

Then add an integration test that composes the components and drives a realistic webhook delivery scenario end to end. Assert the happy path first, then deliberately trigger an endpoint that returns 200 to a malformed payload and silently swallows webhooks that never take effect and assert the system degrades rather than corrupts. Because the components are pure where practical, the integration test stays deterministic rather than flaky.

A webhook delivery suite must also prove the negative: bad input is rejected at the boundary, overload is shed rather than queued indefinitely, and a partial failure does not leave the system in an inconsistent state. These are the tests that catch the regressions a happy-path suite never would.

The entire webhook delivery suite runs with no external service; every dependency is either in-process or stubbed, so every case is reproducible in CI. That is what lets the attempt runner failure-mode assertions live as tests rather than as a manual checklist that drifts out of date.

# run the webhook delivery test suite
pnpm vitest run

Deployment and operations

Run at least two webhook delivery instances behind a load balancer and share state in Redis. Set resource limits so webhook delivery overload surfaces as backpressure rather than failure.

Run at least two stateless instances behind a load balancer, and move any shared state into a shared store so the webhook delivery behavior is global rather than per instance. Size limits so the system sheds load before it runs out of memory, because a webhook delivery system that OOMs under load fails in a way that is hard to reason about.

Drain on deploy: stop accepting new work, let in-flight webhook delivery operations finish, then exit. Because the system holds no state that cannot be reconstructed, a rollout is a matter of cycling instances one at a time while watching the error rate.

When a dependency is unavailable, decide explicitly whether the webhook delivery system fails closed or degrades. Leaving that choice to chance is how a dependency outage becomes a silent data loss, so make the choice visible in code and in a test.

Security considerations

Validate all webhook delivery inputs at the boundary, reject oversized payloads before buffering them, and keep secrets in the environment rather than the repository.

The webhook delivery system is a choke point, so it is the right place to reject malformed input before it reaches backoff scheduler or attempt runner. Validate the shape and the size at the edge, and let the inner components assume what they receive is well-formed.

Never log secrets or full payloads; log enough to localize a failure without recording data that should not be on disk. Treat any external call as untrusted and validate its response before using it, because a webhook delivery system that trusts its inputs blindly is one malformed response away from corruption.

Keep dependencies pinned and run an audit in CI. A transitive dependency with a known vulnerability is still a vulnerability in your webhook delivery system, and the lockfile is what makes the audit meaningful.

Observability

Emit one structured event per webhook delivery operation with the inputs’ shape, the outcome, and a correlation id. Track error rate and latency per component and alert on a step change rather than a single spike.

The correlation id is the thread that ties a user-visible symptom to the webhook delivery decision and to the component log entry for the same operation. Without it, an investigation starts from a timestamp and a guess.

Watch the attempt runner failure rate separately from the overall error rate. A jump in attempt runner failures usually means an endpoint that returns 200 to a malformed payload and silently swallows webhooks that never take effect, not a bug in backoff scheduler, and the two have different owners and different fixes.

Chart the share of webhook delivery operations that shed, queue, or succeed. A change in the mix points at the cause faster than an aggregate latency can, because each class has a different root.

Troubleshooting

An endpoint that returns 200 to a malformed payload and silently swallows webhooks that never take effect. Reproduce with a minimal fixture, then add a guard test so the webhook delivery system rejects the condition loudly.

Performance regresses under webhook delivery load. Profile with the sampling analyzer, locate the hot path in backoff scheduler, and add a targeted benchmark to lock the regression.

Keep the webhook delivery dashboards separate from dependency dashboards. When an operation fails, the correlation id lets you decide in seconds whether the fault is delivery queue, backoff scheduler, or attempt runner, instead of guessing from a timeout.

Limitations and next steps

• Single-region; multi-region replication is out of scope.
• The delivery queue is in-process and not horizontally shared by default.
• No auth layer; assume deployment behind an edge that enforces identity.

The webhook delivery system is single-region by design; extending to multi-region means making delivery queue shareable across regions, which is a real change to the consistency model rather than a config flag. The attempt runner is in-process, which is correct for one instance but must move to a shared store before scaling out.

There is no authentication or authorization layer; the system assumes it is deployed behind an edge that enforces identity. Adding one is straightforward but out of scope for this build, and doing it here would distract from the webhook delivery correctness that is the point.

The natural next step, once the webhook delivery pipeline is trusted in production, is to add a short-TTL cache for idempotent reads and a richer result classifier that exports histograms rather than counters. Both are additive and neither changes the contracts this build establishes.