Never stop coding. Free MIT AI gateway: one endpoint, 375 providers (150+ free), 1,400+ models Kimi, Claude, GPT, Gemini, GLM, DeepSeek, MiniMax. Works with Claude Code, Codex, Cursor, OpenCode, Cline & Copilot. Quota-aware auto-fallback, RTK+Caveman compression saves 15-95% tokens, MCP/A2A, Desktop/PWA. Built by 900+ contributors
TypeScript a2aai-agentsai-gatewayanthropicclaudeclaude-codeclinecodexcopilotcursordeepseekfree-aigeminikimillm-gatewaymcpopenaiopenai-proxyqwentoken-saver
// tests/unit/chat-keepalive-correlation-functional.test.ts
// Route-level keepalive correlation (#14792, follow-up #14850).
//
// /v1/chat/completions passes its request id to withEarlyStreamKeepalive as
// `correlationId` (src/app/api/v1/chat/completions/route.ts). With it, every byte the
// wrapper writes straight to the client is buffered under that id
// (open-sse/utils/earlyKeepaliveByteBuffer.ts) so the handler can merge it into the
// call-log row. Without it, nothing is buffered.
//
// - The first test is the regression guard for that route line: it reads the real
// buffer after the wrapper's first frame and fails when the route stops passing
// `correlationId` (verified red with the line removed, green with it).
// - The second is an end-to-end check that the persisted row carries the keepalive
// bytes. Neither test records bytes on the wrapper's behalf.
//
// The upstream is held open by a promise the test releases, instead of a fixed sleep.
// The only real wait is the route's own 1s keepalive threshold for openai/* models.
import test from "node:test";
import assert from "node:assert/strict";
import fs from "node:fs";
import os from "node:os";
import path from "node:path";
import { CALL_LOG_SETTLE_TIMEOUT_MS } from "../helpers/callLogSettle.ts";
const dataDir = fs.mkdtempSync(path.join(os.tmpdir(), "omniroute-chat-keepalive-func-"));
process.env.DATA_DIR = dataDir;
process.env.REQUIRE_API_KEY = "false";
// The attempt-logging merge runs only when detailed logging is enabled and a
// correlation id is present; chatCore reads the DB-backed setting, seeded per test.
process.env.CALL_LOG_PIPELINE_CAPTURE_STREAM_CHUNKS = "true";
const core = await import("../../src/lib/db/core.ts");
const providersDb = await import("../../src/lib/db/providers.ts");
const settingsDb = await import("../../src/lib/db/settings.ts");
const chatRoute = await import("../../src/app/api/v1/chat/completions/route.ts");
const { takeEarlyKeepaliveBytes } =
await import("../../open-sse/utils/earlyKeepaliveByteBuffer.ts");
const originalFetch = globalThis.fetch;
const UPSTREAM_SSE =
'data: {"id":"chatcmpl-keepalive-func","object":"chat.completion.chunk","created":1,"model":"gpt-4.1","choices":[{"index":0,"delta":{"content":"OK"},"finish_reason":null}]}\n\ndata: [DONE]\n\n';
async function flushBackgroundWork() {
await new Promise((resolve) => setTimeout(resolve, 20));
await new Promise((resolve) => setImmediate(resolve));
}
test.beforeEach(async () => {
globalThis.fetch = originalFetch;
core.resetDbInstance();
fs.rmSync(dataDir, { recursive: true, force: true, maxRetries: 5, retryDelay: 100 });
fs.mkdirSync(dataDir, { recursive: true });
await core.ensureDbInitialized();
await providersDb.createProviderConnection({
provider: "openai",
authType: "apikey",
name: "openai-chat-keepalive-func",
apiKey: "sk-chat-keepalive-func",
isActive: true,
testStatus: "active",
});
});
test.afterEach(async () => {
await flushBackgroundWork();
globalThis.fetch = originalFetch;
});
test.after(async () => {
await flushBackgroundWork();
globalThis.fetch = originalFetch;
core.resetDbInstance();
fs.rmSync(dataDir, { recursive: true, force: true, maxRetries: 5, retryDelay: 100 });
});
/** Upstream that answers only once the test calls `release()`. */
function holdUpstream(): { release: () => void } {
let release: () => void = () => {};
const released = new Promise<void>((resolve) => {
release = resolve;
});
globalThis.fetch = (async () => {
await released;
return new Response(UPSTREAM_SSE, {
status: 200,
headers: { "Content-Type": "text/event-stream" },
});
}) as typeof fetch;
return { release: () => release() };
}
function streamingRequest(correlationId: string): Request {
return new Request("http://localhost/v1/chat/completions", {
method: "POST",
headers: { "Content-Type": "application/json", "X-Correlation-Id": correlationId },
body: JSON.stringify({
model: "openai/gpt-4.1",
messages: [{ role: "user", content: "hi" }],
stream: true,
}),
});
}
test("route buffers its keepalive bytes under the caller's X-Correlation-Id", async () => {
const upstream = holdUpstream();
const correlationId = `chat-keepalive-wiring-${Date.now()}`;
const res = await chatRoute.POST(streamingRequest(correlationId));
assert.equal(res.status, 200);
assert.equal(res.headers.get("x-correlation-id"), correlationId);
// The slow path has committed: the wrapper has written its startup frame to the wire.
const reader = res.body!.getReader();
const first = await reader.read();
assert.match(new TextDecoder().decode(first.value), /chatcmpl-keepalive/);
// Only the route's `correlationId: reqId` makes the wrapper buffer what it wrote.
const buffered = takeEarlyKeepaliveBytes(correlationId).join("");
assert.match(
buffered,
/chatcmpl-keepalive/,
"withEarlyStreamKeepalive must receive the request's correlation id from the route"
);
upstream.release();
await reader.cancel().catch(() => {});
});
test("slow streaming chat request persists keepalive bytes in the call-log row", async () => {
await settingsDb.updateSettings({ call_log_pipeline_enabled: true });
const upstream = holdUpstream();
const correlationId = `chat-keepalive-func-${Date.now()}`;
const res = await chatRoute.POST(streamingRequest(correlationId));
assert.equal(res.status, 200);
const reader = res.body!.getReader();
const decoder = new TextDecoder();
let wire = decoder.decode((await reader.read()).value);
// Startup frame is on the wire; now let the upstream answer and drain the rest so
// the handler finishes and persists the attempt log.
upstream.release();
for (;;) {
const { done, value } = await reader.read();
if (done) break;
wire += decoder.decode(value);
}
assert.match(wire, /chatcmpl-keepalive/, "wire must carry the keepalive startup frame");
assert.match(wire, /chatcmpl-keepalive-func/, "wire must carry the upstream chunk");
// The journal row is keyed on an internal traceId, not the caller correlation id:
// find the recent row carrying ours, then re-read it in full (summary rows lack
// the pipeline payloads). Persistence is async, so poll briefly.
const { getCallLogs, getCallLogById } = await import("../../src/lib/usage/callLogs.ts");
const deadline = Date.now() + CALL_LOG_SETTLE_TIMEOUT_MS;
let row: Record<string, unknown> | null = null;
while (!row && Date.now() < deadline) {
const recent = (await getCallLogs({ limit: 10 })) as Array<Record<string, unknown>>;
const summary = recent.find((r) => JSON.stringify(r).includes(correlationId));
if (summary?.id && summary.detailState === "ready") {
row = (await getCallLogById(summary.id as string)) as Record<string, unknown> | null;
}
if (!row) await new Promise((resolve) => setTimeout(resolve, 50));
}
assert.ok(row, "call log row should be persisted");
const payload = (row.pipelinePayloads ?? row.pipeline ?? {}) as {
streamChunks?: { client?: string[] };
};
const client = (payload.streamChunks?.client ?? []).join("");
assert.match(
client,
/chatcmpl-keepalive/,
"persisted streamChunks.client must contain the keepalive bytes written directly to the wire"
);
});