阅读中文版本

Build a Stream Window Aggregator

Stream Window Aggregator — a complete walkthrough of stream processing, covering event ingest, window manager, aggregate function, and testing.

stream_window_aggregator/
├── src/index.ts
├── src/event-ingest.ts
├── src/window-manager.ts
├── tests/stream_window_aggregator.test.ts
└── package.json

// Event ingest: stream processing core
export function eventingest(input: EventingestInput): EventingestResult {
  // Build the event ingest that anchors stream processing and defines the core data shape.
  return processEventingest(input);
}

// Window manager — handles stream processing Add the window manager so
export class Windowmanager {
  constructor(private opts: WindowmanagerOptions) {}
  async run(ctx: Context): Promise<WindowmanagerOutput> {
    return this.execute(ctx);
  }
}

// Aggregate function: stream processing stage
export const aggregatefunction = (next: Next) => async (req: Request) => {
  const out = await transform(req);
  return next(out);
};

// Late event handler for stream processing
export async function lateeventhandler(stream: AsyncIterable<Chunk>): Promise<Aggregate> {
  let acc = init();
  for await (const c of stream) acc = merge(acc, c);
  return acc;
}

// Watermark tracker: stream processing finalization
export function watermarktracker(state: State): Result {
  validate(state);
  return finalize(state);
}

Project overview

This project builds a stream processing system whose reason for existing is computing tumbling and sliding window aggregates over an unbounded event stream with bounded memory. The temptation in stream processing is to bolt the interesting part onto a working prototype and discover the hard cases in production; this build inverts that order and treats a late event that arrives after its window closed and must either update a closed result or be dropped to a side output as the first thing designed, not the last.

You will implement Event ingest, Window manager, Aggregate function, Late event handler, and Watermark tracker, each as a module with one responsibility and a test that proves it in isolation. The stream processing pipeline is then the wiring that composes them, and the integration test is where the composition is shown to survive a late event that arrives after its window closed and must either update a closed result or be dropped to a side output.

The walkthrough is written for a developer who has used a stream processing 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 event ingest that returns the right shape, a window manager that rejects bad input, an integration test that goes green — so you can stop at any point and still have a stream processing system that works for the cases covered so far.

What you will build

The finished stream processing system exposes five capabilities, each with a focused test:

• Event ingest
• Window manager
• Aggregate function
• Late event handler
• Watermark tracker

Event ingest is the foundation that every later stage assumes; Window manager is where stream processing correctness lives; Aggregate function handles a late event that arrives after its window closed and must either update a closed result or be dropped to a side output; Late event handler makes the system observable; and Watermark tracker is the graceful-degradation layer that keeps overload from becoming corruption.

The build keeps the stream processing 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 Event ingest, Window manager, and Aggregate function, with Late event handler tying them together. Data flows as a set of windows each holding a partial aggregate, closed by a watermark that advances with observed event timestamps.

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 Aggregate function does not need to stand up Event ingest and Window manager to exercise the interesting branch.

The data model is deliberately explicit. A set of windows each holding a partial aggregate, closed by a watermark that advances with observed event timestamps 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 stream processing 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 Aggregate function is designed before the happy path is finished, because a late event that arrives after its window closed and must either update a closed result or be dropped to a side output is the case that distinguishes a stream processing 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 stream processing setup is not blocked by a missing tool:

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

The runtime dependencies are chosen to keep the stream processing datapath legible:

• fastify (^5.0.0) — Stream API
• redis (^4.7.0) — Window state

# setup for stream-window-aggregator
pnpm init
pnpm add fastify redis
pnpm add -D typescript tsx vitest

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

Directory structure

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

stream_window_aggregator/
├── src/index.ts
├── src/event-ingest.ts
├── src/window-manager.ts
├── tests/stream_window_aggregator.test.ts
└── package.json

Phase 1: Event ingest

Build the event ingest that anchors stream processing and defines the core data shape.

// Event ingest: stream processing core
export function eventingest(input: EventingestInput): EventingestResult {
  // Build the event ingest that anchors stream processing and defines the core data shape.
  return processEventingest(input);
}

After this phase the event ingest works in isolation and is covered by a unit test. The event ingest 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 event ingest as the single source of truth for the stream processing data shape. Validate inputs at the boundary and reject early, so a malformed value never reaches Window manager or Aggregate function where its effect would be harder to localize.

Phase 2: Window manager

Add the window manager so stream processing can computing tumbling and sliding correctly.

// Window manager — handles stream processing Add the window manager so
export class Windowmanager {
  constructor(private opts: WindowmanagerOptions) {}
  async run(ctx: Context): Promise<WindowmanagerOutput> {
    return this.execute(ctx);
  }
}

After this phase the window manager works in isolation and is covered by a unit test. This is where stream processing correctness lives: the window manager is what makes computing tumbling and sliding window aggregates over an unbounded event stream with bounded memory true, and its test is the proof.

Keep the window manager 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 window manager about the transformation.

Phase 3: Aggregate function

Wire the aggregate function to handle a late event that without losing data.

// Aggregate function: stream processing stage
export const aggregatefunction = (next: Next) => async (req: Request) => {
  const out = await transform(req);
  return next(out);
};

After this phase the aggregate function works in isolation and is covered by a unit test. The aggregate function is the difference between a stream processing system that works on the happy path and one that survives a late event that arrives after its window closed and must either update a closed result or be dropped to a side output.

Design the failure before the success. Decide what aggregate function does when a late event that arrives after its window closed and must either update a closed result or be dropped to a side output: reject, retry, shed, or degrade — and write that test first. A stream processing system whose failure behavior is unspecified will pick one at random under pressure, and it is usually the wrong one.

Phase 4: Late event handler

Integrate the late event handler to make stream processing observable and testable end to end.

// Late event handler for stream processing
export async function lateeventhandler(stream: AsyncIterable<Chunk>): Promise<Aggregate> {
  let acc = init();
  for await (const c of stream) acc = merge(acc, c);
  return acc;
}

After this phase the late event handler works in isolation and is covered by a unit test. Observability is not a finishing touch; it is how you know the stream processing 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 late event handler is what turns “something is slow” into “the window manager is slow for inputs matching this shape,” which is the difference between an alert and an investigation.

Phase 5: Watermark tracker

Finalize the watermark tracker so the stream processing system degrades gracefully under overload.

// Watermark tracker: stream processing finalization
export function watermarktracker(state: State): Result {
  validate(state);
  return finalize(state);
}

After this phase the watermark tracker works in isolation and is covered by a unit test. The watermark tracker is what keeps overload from becoming corruption: when the stream processing 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 watermark tracker — shed, queue, or backpressure — and make it explicit in code and in a test. Leaving the behavior to chance is how a stream processing traffic spike becomes a data loss incident.

Testing

Each phase has a focused test: event ingest, window manager, and aggregate function are exercised in isolation, then an integration test drives the full stream processing pipeline and asserts the a late event that case is handled.

Start with unit tests that pin each component’s contract: the event ingest returns the documented shape, the window manager transforms correctly, and the aggregate function 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 stream processing scenario end to end. Assert the happy path first, then deliberately trigger a late event that arrives after its window closed and must either update a closed result or be dropped to a side output and assert the system degrades rather than corrupts. Because the components are pure where practical, the integration test stays deterministic rather than flaky.

A stream processing 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 stream processing 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 aggregate function failure-mode assertions live as tests rather than as a manual checklist that drifts out of date.

# run the stream processing test suite
pnpm vitest run

Deployment and operations

Run at least two stream processing instances behind a load balancer and share state in Redis. Set resource limits so stream processing 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 stream processing behavior is global rather than per instance. Size limits so the system sheds load before it runs out of memory, because a stream processing 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 stream processing 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 stream processing 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 stream processing inputs at the boundary, reject oversized payloads before buffering them, and keep secrets in the environment rather than the repository.

The stream processing system is a choke point, so it is the right place to reject malformed input before it reaches window manager or aggregate function. 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 stream processing 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 stream processing system, and the lockfile is what makes the audit meaningful.

Observability

Emit one structured event per stream processing 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 stream processing decision and to the component log entry for the same operation. Without it, an investigation starts from a timestamp and a guess.

Watch the aggregate function failure rate separately from the overall error rate. A jump in aggregate function failures usually means a late event that arrives after its window closed and must either update a closed result or be dropped to a side output, not a bug in window manager, and the two have different owners and different fixes.

Chart the share of stream processing 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

A late event that arrives after its window closed and must either update a closed result or be dropped to a side output. Reproduce with a minimal fixture, then add a guard test so the stream processing system rejects the condition loudly.

Performance regresses under stream processing load. Profile with the sampling analyzer, locate the hot path in window manager, and add a targeted benchmark to lock the regression.

Keep the stream processing dashboards separate from dependency dashboards. When an operation fails, the correlation id lets you decide in seconds whether the fault is event ingest, window manager, or aggregate function, instead of guessing from a timeout.

Limitations and next steps

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

The stream processing system is single-region by design; extending to multi-region means making event ingest shareable across regions, which is a real change to the consistency model rather than a config flag. The aggregate function 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 stream processing correctness that is the point.

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