阅读中文版本

Build a Raft Leader Election

Raft Leader Election — a complete walkthrough of consensus leadership, covering term tracker, vote requester, heartbeat sender, and testing.

leader_election_raft/
├── src/index.ts
├── src/term-tracker.ts
├── src/vote-requester.ts
├── tests/leader_election_raft.test.ts
└── package.json

// Term tracker: consensus leadership core
export function termtracker(input: TermtrackerInput): TermtrackerResult {
  // Build the term tracker that anchors consensus leadership and defines the core data shape.
  return processTermtracker(input);
}

// Vote requester — handles consensus leadership Add the vote requester so
export class Voterequester {
  constructor(private opts: VoterequesterOptions) {}
  async run(ctx: Context): Promise<VoterequesterOutput> {
    return this.execute(ctx);
  }
}

// Heartbeat sender: consensus leadership stage
export const heartbeatsender = (next: Next) => async (req: Request) => {
  const out = await transform(req);
  return next(out);
};

// Log replicator for consensus leadership
export async function logreplicator(stream: AsyncIterable<Chunk>): Promise<Aggregate> {
  let acc = init();
  for await (const c of stream) acc = merge(acc, c);
  return acc;
}

// State machine: consensus leadership finalization
export function statemachine(state: State): Result {
  validate(state);
  return finalize(state);
}

Project overview

This project builds a consensus leadership system whose reason for existing is electing a single leader that survives network partitions and node restarts without split brain. The temptation in consensus leadership 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 partitioned minority that cannot reach a majority and must stay inert to avoid electing a competing leader as the first thing designed, not the last.

You will implement Term tracker, Vote requester, Heartbeat sender, Log replicator, and State machine, each as a module with one responsibility and a test that proves it in isolation. The consensus leadership pipeline is then the wiring that composes them, and the integration test is where the composition is shown to survive a partitioned minority that cannot reach a majority and must stay inert to avoid electing a competing leader.

The walkthrough is written for a developer who has used a consensus leadership 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 term tracker that returns the right shape, a vote requester that rejects bad input, an integration test that goes green — so you can stop at any point and still have a consensus leadership system that works for the cases covered so far.

What you will build

The finished consensus leadership system exposes five capabilities, each with a focused test:

• Term tracker
• Vote requester
• Heartbeat sender
• Log replicator
• State machine

Term tracker is the foundation that every later stage assumes; Vote requester is where consensus leadership correctness lives; Heartbeat sender handles a partitioned minority that cannot reach a majority and must stay inert to avoid electing a competing leader; Log replicator makes the system observable; and State machine is the graceful-degradation layer that keeps overload from becoming corruption.

The build keeps the consensus leadership 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 Term tracker, Vote requester, and Heartbeat sender, with Log replicator tying them together. Data flows as a monotonic term, a voted-for field, and a log of committed entries that a leader replicates to a majority before applying.

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 Heartbeat sender does not need to stand up Term tracker and Vote requester to exercise the interesting branch.

The data model is deliberately explicit. A monotonic term, a voted-for field, and a log of committed entries that a leader replicates to a majority before applying 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 consensus leadership 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 Heartbeat sender is designed before the happy path is finished, because a partitioned minority that cannot reach a majority and must stay inert to avoid electing a competing leader is the case that distinguishes a consensus leadership 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 consensus leadership setup is not blocked by a missing tool:

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

The runtime dependencies are chosen to keep the consensus leadership datapath legible:

• fastify (^5.0.0) — Peer RPC
• pino (^9.0.0) — Consensus logging

# setup for leader-election-raft
pnpm init
pnpm add fastify pino
pnpm add -D typescript tsx vitest

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

Directory structure

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

leader_election_raft/
├── src/index.ts
├── src/term-tracker.ts
├── src/vote-requester.ts
├── tests/leader_election_raft.test.ts
└── package.json

Phase 1: Term tracker

Build the term tracker that anchors consensus leadership and defines the core data shape.

// Term tracker: consensus leadership core
export function termtracker(input: TermtrackerInput): TermtrackerResult {
  // Build the term tracker that anchors consensus leadership and defines the core data shape.
  return processTermtracker(input);
}

After this phase the term tracker works in isolation and is covered by a unit test. The term tracker 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 term tracker as the single source of truth for the consensus leadership data shape. Validate inputs at the boundary and reject early, so a malformed value never reaches Vote requester or Heartbeat sender where its effect would be harder to localize.

Phase 2: Vote requester

Add the vote requester so consensus leadership can electing a single leader correctly.

// Vote requester — handles consensus leadership Add the vote requester so
export class Voterequester {
  constructor(private opts: VoterequesterOptions) {}
  async run(ctx: Context): Promise<VoterequesterOutput> {
    return this.execute(ctx);
  }
}

After this phase the vote requester works in isolation and is covered by a unit test. This is where consensus leadership correctness lives: the vote requester is what makes electing a single leader that survives network partitions and node restarts without split brain true, and its test is the proof.

Keep the vote requester 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 vote requester about the transformation.

Phase 3: Heartbeat sender

Wire the heartbeat sender to handle a partitioned minority that without losing data.

// Heartbeat sender: consensus leadership stage
export const heartbeatsender = (next: Next) => async (req: Request) => {
  const out = await transform(req);
  return next(out);
};

After this phase the heartbeat sender works in isolation and is covered by a unit test. The heartbeat sender is the difference between a consensus leadership system that works on the happy path and one that survives a partitioned minority that cannot reach a majority and must stay inert to avoid electing a competing leader.

Design the failure before the success. Decide what heartbeat sender does when a partitioned minority that cannot reach a majority and must stay inert to avoid electing a competing leader: reject, retry, shed, or degrade — and write that test first. A consensus leadership system whose failure behavior is unspecified will pick one at random under pressure, and it is usually the wrong one.

Phase 4: Log replicator

Integrate the log replicator to make consensus leadership observable and testable end to end.

// Log replicator for consensus leadership
export async function logreplicator(stream: AsyncIterable<Chunk>): Promise<Aggregate> {
  let acc = init();
  for await (const c of stream) acc = merge(acc, c);
  return acc;
}

After this phase the log replicator works in isolation and is covered by a unit test. Observability is not a finishing touch; it is how you know the consensus leadership 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 log replicator is what turns “something is slow” into “the vote requester is slow for inputs matching this shape,” which is the difference between an alert and an investigation.

Phase 5: State machine

Finalize the state machine so the consensus leadership system degrades gracefully under overload.

// State machine: consensus leadership finalization
export function statemachine(state: State): Result {
  validate(state);
  return finalize(state);
}

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

Testing

Each phase has a focused test: term tracker, vote requester, and heartbeat sender are exercised in isolation, then an integration test drives the full consensus leadership pipeline and asserts the a partitioned minority that case is handled.

Start with unit tests that pin each component’s contract: the term tracker returns the documented shape, the vote requester transforms correctly, and the heartbeat sender 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 consensus leadership scenario end to end. Assert the happy path first, then deliberately trigger a partitioned minority that cannot reach a majority and must stay inert to avoid electing a competing leader and assert the system degrades rather than corrupts. Because the components are pure where practical, the integration test stays deterministic rather than flaky.

A consensus leadership 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 consensus leadership 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 heartbeat sender failure-mode assertions live as tests rather than as a manual checklist that drifts out of date.

# run the consensus leadership test suite
pnpm vitest run

Deployment and operations

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

The consensus leadership system is a choke point, so it is the right place to reject malformed input before it reaches vote requester or heartbeat sender. 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 consensus leadership 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 consensus leadership system, and the lockfile is what makes the audit meaningful.

Observability

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

Watch the heartbeat sender failure rate separately from the overall error rate. A jump in heartbeat sender failures usually means a partitioned minority that cannot reach a majority and must stay inert to avoid electing a competing leader, not a bug in vote requester, and the two have different owners and different fixes.

Chart the share of consensus leadership 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 partitioned minority that cannot reach a majority and must stay inert to avoid electing a competing leader. Reproduce with a minimal fixture, then add a guard test so the consensus leadership system rejects the condition loudly.

Performance regresses under consensus leadership load. Profile with the sampling analyzer, locate the hot path in vote requester, and add a targeted benchmark to lock the regression.

Keep the consensus leadership dashboards separate from dependency dashboards. When an operation fails, the correlation id lets you decide in seconds whether the fault is term tracker, vote requester, or heartbeat sender, instead of guessing from a timeout.

Limitations and next steps

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

The consensus leadership system is single-region by design; extending to multi-region means making term tracker shareable across regions, which is a real change to the consistency model rather than a config flag. The heartbeat sender 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 consensus leadership correctness that is the point.

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