fix(routes): keep the 12 s budget from arrival and name the outage to voice

Starting the upstream timeout after the outbound queue let a request run
up to about 19 s, past the Directions layer's 15 s; the budget again runs
from arrival, queue wait included, and a budget spent before the request
leaves the queue answers 429 "routing busy" rather than an outage.

plan_route threw "Routing HTTP 502" for an outage, and voice passes only
the message on; it now throws "Routing service unavailable (HTTP 502)".

Co-Authored-By: Claude Opus 5.5 <[email protected]>

changelog.d/1509-route-outage-unavailable.md

@@ -1,4 +1,4 @@
-- A routing service outage (an error status, no answer or a non-JSON body)
-  is reported as "routing service unavailable" instead of "no route found",
-  so voice route planning and the Directions layer no longer blame the
-  request (quanticsoul4772, #1509).
+- When the routing service is down (an error status, no answer or a non-JSON
+  body), voice route planning hears "Routing service unavailable" and the
+  Directions layer shows "Routing unavailable" instead of "no route found",
+  so neither blames the request (quanticsoul4772, #1509).

docs/CURRENT-STATE.md

@@ -4376,13 +4376,15 @@ service's usage policy states; past a bounded queue the answer is a 429 with a
 upstream host is pinned and a redirect is refused rather than followed, and a
 rate limit from the routing service is reported as one instead of as "no route
 found". So is an outage: any failure status but 400 (a 5xx, a 403 block, a
-404), no answer (a 12 s timeout counted from when the request leaves, not
-while it waits for an outbound slot; a network error) or a body that is not
-JSON answers
-`502 routing service unavailable`, which `plan_route`
-reports as `unavailable` and the layer as `Routing unavailable (HTTP 502)`;
-only OSRM's own refusal (a 400, or a 200 without a route) reads "no route
-found".
+404), no answer (timeout, network error) or a body that is not JSON answers
+`502 routing service unavailable`. `plan_route` throws `Routing service
+unavailable (HTTP 502)`, which is what voice hears, and the layer shows
+`Routing unavailable (HTTP 502)`; only OSRM's own refusal (a 400, or a 200
+without a route) reads "no route found". The 12 s timeout runs from the
+request's arrival, queue wait included; a budget spent before the request
+leaves the queue answers 429 "routing busy". A request that waited in the
+queue gives OSRM the rest of the 12 s, so a slow answer then still reads as
+unavailable.
 `src/data/routeSteps.js` turns OSRM maneuver type / modifier / exit /
 road name into one plain-English sentence per decision, folding
 `exit roundabout` steps into the roundabout they leave, and caps one route at

server/providers/places/routes.js

@@ -178,14 +178,16 @@ async function fetchRoute({
     `${base.replace(/\/$/, '')}/route/v1/${osrmProfile}/${coords}` +
     `?overview=full&geometries=geojson&alternatives=false&steps=${withSteps ? 'true' : 'false'}`;
   const controller = new AbortController();
-  let timer;
+  // One 12 s budget from arrival, queue wait included, so callers sized to it
+  // (the Directions layer gives up at 15 s) hear back in time.
+  const timer = setTimeout(() => controller.abort(), 12000);
   let osrm;
   try {
     const upstreamRes = await throughUpstreamGate(async () => {
-      // The 12 s budget starts when the request leaves, not while it waits
-      // for an outbound slot, so a wait in our own queue is never reported
-      // as the routing service being unavailable.
-      timer = setTimeout(() => controller.abort(), 12000);
+      // A budget spent waiting in our own queue is our backlog, not the
+      // routing service failing.
+      if (controller.signal.aborted)
+        throw new RouteBusyError(_upstreamIntervalMs);
       try {
         return await fetchImpl(upstream, {
           signal: controller.signal,

src/tooling/placeProviders.test.mjs

@@ -311,15 +311,16 @@ test('callers sharing one failed upstream call all hear the outage, and it is no
   assert.equal(calls, 2, 'the next request asks the routing service again');
 });
 
-test('a wait for an outbound slot does not count toward the upstream timeout', async (t) => {
+test('a budget spent in the outbound queue is reported as busy, not as an outage', async (t) => {
   _resetRouteUpstreamForTest();
-  // The second request waits 13 s for its slot, longer than the 12 s budget.
+  // The second request waits 13 s for its slot, past the 12 s budget.
   _setRouteUpstreamIntervalForTest(13_000);
   t.after(() => _resetRouteUpstreamForTest());
   t.mock.timers.enable({ apis: ['setTimeout', 'Date'], now: Date.now() });
   const request = install(installRouteMiddleware);
-  t.mock.method(globalThis, 'fetch', async (_url, { signal }) => {
-    signal.throwIfAborted();
+  let calls = 0;
+  t.mock.method(globalThis, 'fetch', async () => {
+    calls += 1;
     return Response.json(osrmRouteWithSteps(2));
   });
   const first = request(
@@ -333,8 +334,9 @@ test('a wait for an outbound slot does not count toward the upstream timeout', a
   assert.equal((await first).body.ok, true);
   t.mock.timers.tick(13_000);
   const result = await second;
-  assert.equal(result.statusCode, 200);
-  assert.equal(result.body.ok, true);
+  assert.equal(result.statusCode, 429);
+  assert.match(result.body.error, /routing busy/);
+  assert.equal(calls, 1, 'the spent request never reaches the routing service');
 });
 
 /** One OSRM route with `count` maneuvers, for the step tests. */

src/tools/places.js

@@ -190,9 +190,13 @@ export function createRouteService({
         /* the status below is authoritative */
       }
       // A rate limit or an outage (5xx) is the routing service's failure, not
-      // a missing route, whatever the body says.
-      if (response.status === 429 || response.status >= 500)
-        throw httpFailure(response, 'Routing');
+      // a missing route, whatever the body says. Voice hears this message.
+      if (response.status === 429) throw httpFailure(response, 'Routing');
+      if (response.status >= 500)
+        throw new ToolError(
+          'unavailable',
+          `Routing service unavailable (HTTP ${response.status})`,
+        );
       if (payload?.ok === false || response.status === 400)
         return { ok: false, error: payload?.error || 'no route found' };
       if (!response.ok) throw httpFailure(response, 'Routing');

src/tools/queries/places.test.mjs

@@ -283,6 +283,24 @@ test('a routing outage is reported as unavailable, not as invalid arguments', as
     assert.equal(result.structuredContent.error, 'unavailable');
     assert.doesNotMatch(result.content[0].text, /No walk route/);
   }
+  // Voice passes only the message on, so it names the outage itself.
+  const outage = createRouteService({
+    fetchImpl: async () =>
+      Response.json(
+        { ok: false, error: 'routing service unavailable' },
+        { status: 502 },
+      ),
+  });
+  await assert.rejects(
+    outage.route(
+      [
+        { lat: 30, lon: -97 },
+        { lat: 30.01, lon: -97.01 },
+      ],
+      'foot',
+    ),
+    { message: 'Routing service unavailable (HTTP 502)' },
+  );
 });
 
 test('place answers say when the search limits may have left places out', async () => {