test relay routing fallback headers (#5526)

Integrated into release/v3.8.42 (round 3). Relay fallback header extraction + tests (drift-shed: dependabot #5415 commit dropped).

src/app/api/v1/relay/chat/completions/route.ts

@@ -20,6 +20,7 @@ import {
 } from "./relaySecurity";
 import {
   getBifrostRoutingConfig,
+  getRoutingFallbackHeader,
   resolveRelayRoutingBackend,
   shouldTryBifrost,
   type BifrostRoutingConfig,
@@ -335,8 +336,9 @@ export async function POST(request: Request) {
     const newHeaders = new Headers(response.headers);
     newHeaders.set("X-Relay-Token", token.tokenPrefix + "...");
     newHeaders.set("X-Routing-Backend", "ts");
-    if (backend === "auto" && bifrostConfig?.enabled) {
-      newHeaders.set("X-Routing-Fallback", "bifrost");
+    const routingFallback = getRoutingFallbackHeader(backend, bifrostConfig);
+    if (routingFallback) {
+      newHeaders.set("X-Routing-Fallback", routingFallback);
     }
 
     return new Response(response.body, {

src/app/api/v1/relay/chat/completions/routingBackend.ts

@@ -10,7 +10,9 @@ export interface BifrostRoutingConfig {
   enabled: boolean;
 }
 
-export function getBifrostRoutingConfig(env: NodeJS.ProcessEnv = process.env): BifrostRoutingConfig | null {
+export function getBifrostRoutingConfig(
+  env: NodeJS.ProcessEnv = process.env
+): BifrostRoutingConfig | null {
   const baseUrl = env.BIFROST_BASE_URL?.replace(/\/$/, "");
   if (!baseUrl) return null;
   const timeoutMs = Number.parseInt(env.BIFROST_TIMEOUT_MS || "", 10);
@@ -24,7 +26,9 @@ export function getBifrostRoutingConfig(env: NodeJS.ProcessEnv = process.env): B
   };
 }
 
-export function resolveRelayRoutingBackend(env: NodeJS.ProcessEnv = process.env): RelayRoutingBackend {
+export function resolveRelayRoutingBackend(
+  env: NodeJS.ProcessEnv = process.env
+): RelayRoutingBackend {
   const configured = env.OMNIROUTE_RELAY_BACKEND || env.RELAY_ROUTING_BACKEND;
   if (configured && VALID_BACKENDS.has(configured as RelayRoutingBackend)) {
     return configured as RelayRoutingBackend;
@@ -39,3 +43,10 @@ export function shouldTryBifrost(
 ): config is BifrostRoutingConfig {
   return Boolean(config?.enabled && backend !== "ts");
 }
+
+export function getRoutingFallbackHeader(
+  backend: RelayRoutingBackend,
+  config: BifrostRoutingConfig | null
+): "bifrost" | undefined {
+  return backend === "auto" && config?.enabled ? "bifrost" : undefined;
+}

tests/unit/api/v1/relay-routing-backend.test.ts

@@ -2,6 +2,7 @@ import { test } from "node:test";
 import assert from "node:assert/strict";
 import {
   getBifrostRoutingConfig,
+  getRoutingFallbackHeader,
   resolveRelayRoutingBackend,
   shouldTryBifrost,
 } from "../../../../src/app/api/v1/relay/chat/completions/routingBackend.ts";
@@ -66,3 +67,34 @@ test("relay routing backend falls back on invalid timeout values", () => {
 
   assert.equal(getBifrostRoutingConfig(env)?.timeoutMs, 30000);
 });
+
+test("relay routing backend exposes TS fallback header only for enabled auto bifrost fallback", () => {
+  const config = getBifrostRoutingConfig({
+    BIFROST_BASE_URL: "http://127.0.0.1:8080",
+  });
+
+  assert.equal(getRoutingFallbackHeader("auto", config), "bifrost");
+  assert.equal(getRoutingFallbackHeader("ts", config), undefined);
+  assert.equal(getRoutingFallbackHeader("bifrost", config), undefined);
+  assert.equal(
+    getRoutingFallbackHeader(
+      "auto",
+      getBifrostRoutingConfig({
+        BIFROST_BASE_URL: "http://127.0.0.1:8080",
+        BIFROST_ENABLED: "0",
+      })
+    ),
+    undefined
+  );
+  assert.equal(getRoutingFallbackHeader("auto", null), undefined);
+});
+
+test("relay routing backend keeps strict bifrost failures out of auto fallback accounting", () => {
+  const config = getBifrostRoutingConfig({
+    BIFROST_BASE_URL: "http://127.0.0.1:8080",
+  });
+
+  assert.equal(shouldTryBifrost("bifrost", config), true);
+  assert.equal(getRoutingFallbackHeader("bifrost", config), undefined);
+  assert.equal(resolveRelayRoutingBackend({ OMNIROUTE_RELAY_BACKEND: "bifrost" }), "bifrost");
+});