import { describe, it, expect, beforeAll, afterAll, beforeEach } from "vitest";
import {
startServer,
stopServer,
getTestClient,
resetRateLimit,
getBaseUrl,
} from "./setup";
import { APIError } from "openai";
describe("Rate Limiting", () => {
describe("after_n mode", () => {
beforeAll(async () => {
await startServer({
RATE_LIMIT_ENABLED: "true",
RATE_LIMIT_MODE: "after_n",
RATE_LIMIT_AFTER_N: "3",
});
});
afterAll(async () => {
await stopServer();
});
beforeEach(async () => {
await resetRateLimit();
});
it("should allow first N requests and then rate limit", async () => {
const client = getTestClient();
const messages = [{ role: "user" as const, content: "Hello" }];
// First 3 requests should succeed
for (let i = 0; i < 3; i++) {
const response = await client.chat.completions.create({
model: "gpt-4o",
messages,
});
expect(response.choices[0].message.content).toBeTruthy();
}
// 4th request should be rate limited
try {
await client.chat.completions.create({
model: "gpt-4o",
messages,
});
expect.fail("Should have thrown rate limit error");
} catch (error) {
expect(error).toBeInstanceOf(APIError);
const apiError = error as APIError;
expect(apiError.status).toBe(429);
}
});
it("should include rate limit headers", async () => {
const baseUrl = getBaseUrl();
const response = await fetch(`${baseUrl}/v1/chat/completions`, {
method: "POST",
headers: {
"Content-Type": "application/json",
Authorization: "Bearer test-key",
},
body: JSON.stringify({
model: "gpt-4o",
messages: [{ role: "user", content: "Hello" }],
}),
});
expect(response.headers.get("x-ratelimit-limit-requests")).toBeTruthy();
expect(
response.headers.get("x-ratelimit-remaining-requests"),
).toBeTruthy();
});
it("should include retry-after and retry-after-ms headers on 429", async () => {
const baseUrl = getBaseUrl();
// Exhaust the rate limit (3 requests)
for (let i = 0; i < 3; i++) {
await fetch(`${baseUrl}/v1/chat/completions`, {
method: "POST",
headers: {
"Content-Type": "application/json",
Authorization: "Bearer test-key",
},
body: JSON.stringify({
model: "gpt-4o",
messages: [{ role: "user", content: "Hello" }],
}),
});
}
// 4th request should be rate limited with proper headers
const response = await fetch(`${baseUrl}/v1/chat/completions`, {
method: "POST",
headers: {
"Content-Type": "application/json",
Authorization: "Bearer test-key",
},
body: JSON.stringify({
model: "gpt-4o",
messages: [{ role: "user", content: "Hello" }],
}),
});
expect(response.status).toBe(429);
expect(response.headers.get("retry-after")).toBeTruthy();
expect(response.headers.get("retry-after-ms")).toBeTruthy();
// Verify retry-after-ms is a valid number
const retryAfterMs = parseInt(
response.headers.get("retry-after-ms")!,
10,
);
expect(retryAfterMs).toBeGreaterThan(0);
});
it("should return proper error format on rate limit", async () => {
const client = getTestClient();
const messages = [{ role: "user" as const, content: "Hello" }];
// Exhaust the limit
for (let i = 0; i < 3; i++) {
await client.chat.completions.create({ model: "gpt-4o", messages });
}
try {
await client.chat.completions.create({ model: "gpt-4o", messages });
expect.fail("Should have thrown");
} catch (error) {
expect(error).toBeInstanceOf(APIError);
const apiError = error as APIError;
expect(apiError.status).toBe(429);
expect(apiError.message).toContain("Rate limit");
}
});
});
describe("always mode", () => {
beforeAll(async () => {
await startServer({
RATE_LIMIT_ENABLED: "true",
RATE_LIMIT_MODE: "always",
});
});
afterAll(async () => {
await stopServer();
});
it("should rate limit every request", async () => {
const client = getTestClient();
try {
await client.chat.completions.create({
model: "gpt-4o",
messages: [{ role: "user", content: "Hello" }],
});
expect.fail("Should have thrown");
} catch (error) {
expect(error).toBeInstanceOf(APIError);
expect((error as APIError).status).toBe(429);
}
});
});
describe("disabled", () => {
beforeAll(async () => {
await startServer({
RATE_LIMIT_ENABLED: "false",
});
});
afterAll(async () => {
await stopServer();
});
it("should not rate limit when disabled", async () => {
const client = getTestClient();
const messages = [{ role: "user" as const, content: "Hello" }];
// Should be able to make many requests
for (let i = 0; i < 10; i++) {
const response = await client.chat.completions.create({
model: "gpt-4o",
messages,
});
expect(response.choices[0].message.content).toBeTruthy();
}
});
});
});
