Files
LocalAI/core/backend/ctx_propagation_test.go
LocalAI [bot] 06e777b75e feat(distributed): gated X-LocalAI-Node response header (middleware + wrapper) (#9976)
* feat(distributed): add per-request node ID context holder

Introduce pkg/distributedhdr, a leaf package carrying a per-request
*atomic.Value holder for the picked worker node ID from the
SmartRouter (core/services/nodes) up to the HTTP response writer
wrapper (core/http/middleware). Avoids the import cycle that a shared
key in either consumer would create.

Exposes NewHolder, WithHolder, Holder, Stamp, Load, Inherit. The
holder is atomic.Value so cross-goroutine publish from the router to
the response writer wrapper is race-clean.

Assisted-by: Claude:claude-opus-4-7[1m]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* feat(distributed): add ExposeNodeHeader middleware + response writer wrapper

New ApplicationConfig.ExposeNodeHeader bool + --expose-node-header CLI
flag / LOCALAI_EXPOSE_NODE_HEADER env var (default off; the node ID
reveals internal topology and is opt-in).

The middleware creates a per-request *atomic.Value holder, attaches it
to c.Request().Context() via distributedhdr.WithHolder, and wraps
c.Response().Writer with a custom http.ResponseWriter that sets the
X-LocalAI-Node header on first Write / WriteHeader / Flush by reading
the holder. Implements http.Flusher, http.Hijacker, Unwrap so it
composes cleanly with Echo and http.NewResponseController.

request.go propagates the holder onto derived contexts via
distributedhdr.Inherit so the holder survives the correlation-ID
context replacement.

Unit + race-clean concurrency + integration specs.

Assisted-by: Claude:claude-opus-4-7[1m]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* feat(distributed): stamp node ID in router and wire middleware to inference routes

ModelRouterAdapter.Route stamps the picked node ID into the
per-request holder via distributedhdr.Stamp(ctx, result.Node.ID) right
after replica selection.

Wire ExposeNodeHeader middleware to:
- OpenAI chat/completion/embeddings + audio transcriptions/speech + image generations/inpainting
- Anthropic /v1/messages
- Ollama /api/chat, /api/generate, /api/embed, /api/embeddings
- Jina /v1/rerank
- LocalAI /v1/vad

The middleware's wrapper reads the holder on first byte and sets the
X-LocalAI-Node response header before delegating to the underlying
writer. Per-request scope means no race under concurrent multi-replica
routing.

Assisted-by: Claude:claude-opus-4-7[1m]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* fix(distributed): thread request context through backend Load + cover ctx propagation

Five non-OpenAI backend helpers were silently using app.Context instead
of the request context for the gRPC backend call: transcription, TTS,
image generation, rerank, VAD. Effect: distributedhdr.Stamp in the
router callback was a silent no-op for these paths, AND client
cancellation didn't propagate to in-flight inference.

Thread c.Request().Context() (or the equivalent input.Context after
the request middleware has installed the correlation-ID derived
context) through each helper and into ModelOptions via
model.WithContext(ctx). ImageGeneration's signature gains a leading
ctx parameter; in-tree callers (openai image, openai inpainting,
openai inpainting_test) are updated to match.

ModelEmbedding gains a leading ctx parameter for the same reason; the
openai and ollama embedding handlers pass the request context through.

chat_stream_workers.go defers the initial role=assistant chunk
emission until the first token callback so the wrapper's lazy
X-LocalAI-Node lookup against the loader runs AFTER ml.Load has
stamped the per-modelID node ID; semantically identical for clients
(role still arrives before any text).

Regression test core/backend/ctx_propagation_test.go pins ctx
propagation for all five helpers.

Docs updated to enumerate the full endpoint coverage of the
--expose-node-header flag.

Assisted-by: Claude:claude-opus-4-7[1m]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

---------

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Co-authored-by: Ettore Di Giacinto <mudler@localai.io>
2026-05-25 10:51:48 +02:00

170 lines
6.6 KiB
Go

package backend_test
// Regression spec for X-LocalAI-Node coverage on audio/image/TTS/rerank/VAD.
//
// The X-LocalAI-Node middleware (core/http/middleware.ExposeNodeHeader)
// works end-to-end only if the per-request holder attached to the HTTP
// request context reaches the SmartRouter via ml.Load(opts...). The chain
// is:
//
// handler -> backend.Foo(ctx, ...) -> ModelOptions(cfg, app, WithContext(ctx))
// -> ml.Load(opts...) -> grpcModel(..., o.context) -> modelRouter(ctx, ...)
// -> SmartRouter -> distributedhdr.Stamp(ctx, nodeID)
//
// If any backend helper drops `ctx` and lets ModelOptions fall back to the
// app context, the router never sees the per-request holder and the
// header silently stays empty for that endpoint. These specs pin the
// request-context-reaches-router contract for the five backend helpers
// that were previously dropping ctx between the handler and Load.
import (
"context"
"sync/atomic"
"github.com/mudler/LocalAI/core/backend"
"github.com/mudler/LocalAI/core/config"
"github.com/mudler/LocalAI/core/schema"
pbproto "github.com/mudler/LocalAI/pkg/grpc/proto"
"github.com/mudler/LocalAI/pkg/distributedhdr"
"github.com/mudler/LocalAI/pkg/model"
"github.com/mudler/LocalAI/pkg/system"
. "github.com/onsi/ginkgo/v2"
. "github.com/onsi/gomega"
)
// newCapturingLoader returns a ModelLoader wired with a stub model router
// that captures the context it receives and then short-circuits with a
// sentinel error. The router callback is the exact seam where the
// SmartRouter would call distributedhdr.Stamp in production, so observing
// the holder here is equivalent to observing it at the real router.
func newCapturingLoader() (*model.ModelLoader, *atomic.Value, func() context.Context) {
loader := model.NewModelLoader(&system.SystemState{})
var captured atomic.Value
loader.SetModelRouter(func(ctx context.Context, _ string, _, _, _ string, _ *pbproto.ModelOptions, _ bool) (*model.Model, error) {
captured.Store(ctx)
// Return an error so the backend short-circuits before trying to
// dial gRPC. We only care about the context-arrival contract.
return nil, errRouterShortCircuit
})
get := func() context.Context {
v, _ := captured.Load().(context.Context)
return v
}
return loader, &captured, get
}
var errRouterShortCircuit = sentinelErr("router short-circuit (test)")
type sentinelErr string
func (s sentinelErr) Error() string { return string(s) }
func newAppCfg() *config.ApplicationConfig {
return config.NewApplicationConfig(config.WithSystemState(&system.SystemState{}))
}
func newModelCfg() config.ModelConfig {
threads := 1
cfg := config.ModelConfig{
Name: "test-model",
Backend: "stub-backend",
Threads: &threads,
}
cfg.Model = "test.bin"
return cfg
}
var _ = Describe("X-LocalAI-Node ctx propagation contract", func() {
const fakeNodeID = "node-ctx-propagation-7"
var (
appCfg *config.ApplicationConfig
modelCfg config.ModelConfig
loader *model.ModelLoader
routerCtxOf func() context.Context
holder *atomic.Value
reqCtx context.Context
)
BeforeEach(func() {
appCfg = newAppCfg()
modelCfg = newModelCfg()
loader, _, routerCtxOf = newCapturingLoader()
holder = distributedhdr.NewHolder()
reqCtx = distributedhdr.WithHolder(context.Background(), holder)
})
// stampViaRouterCtx asserts the captured router context carries the
// SAME holder that was attached to the request. We verify by stamping
// through the router-side ctx and observing the value via the
// request-side holder; if the holders were different objects the load
// would return "".
stampViaRouterCtx := func() {
routerCtx := routerCtxOf()
Expect(routerCtx).ToNot(BeNil(), "router callback must have been invoked")
distributedhdr.Stamp(routerCtx, fakeNodeID)
Expect(distributedhdr.Load(holder)).To(Equal(fakeNodeID),
"stamp via router-side ctx must be observable via the request-side holder")
}
It("Rerank forwards the request context to the SmartRouter", func() {
_, err := backend.Rerank(reqCtx, &pbproto.RerankRequest{Query: "q"}, loader, appCfg, modelCfg)
Expect(err).To(HaveOccurred())
Expect(err.Error()).To(ContainSubstring("router short-circuit (test)"))
stampViaRouterCtx()
})
It("VAD forwards the request context to the SmartRouter", func() {
_, err := backend.VAD(&schema.VADRequest{}, reqCtx, loader, appCfg, modelCfg)
Expect(err).To(HaveOccurred())
Expect(err.Error()).To(ContainSubstring("router short-circuit (test)"))
stampViaRouterCtx()
})
It("ModelTTS forwards the request context to the SmartRouter", func() {
_, _, err := backend.ModelTTS(reqCtx, "hello", "", "", loader, appCfg, modelCfg)
Expect(err).To(HaveOccurred())
Expect(err.Error()).To(ContainSubstring("router short-circuit (test)"))
stampViaRouterCtx()
})
It("ModelTTSStream forwards the request context to the SmartRouter", func() {
err := backend.ModelTTSStream(reqCtx, "hello", "", "", loader, appCfg, modelCfg, func([]byte) error { return nil })
Expect(err).To(HaveOccurred())
Expect(err.Error()).To(ContainSubstring("router short-circuit (test)"))
stampViaRouterCtx()
})
It("ModelTranscriptionWithOptions forwards the request context to the SmartRouter", func() {
_, err := backend.ModelTranscriptionWithOptions(reqCtx, backend.TranscriptionRequest{Audio: "x.wav"}, loader, modelCfg, appCfg)
Expect(err).To(HaveOccurred())
Expect(err.Error()).To(ContainSubstring("router short-circuit (test)"))
stampViaRouterCtx()
})
It("ModelTranscriptionStream forwards the request context to the SmartRouter", func() {
err := backend.ModelTranscriptionStream(reqCtx, backend.TranscriptionRequest{Audio: "x.wav"}, loader, modelCfg, appCfg, func(backend.TranscriptionStreamChunk) {})
Expect(err).To(HaveOccurred())
Expect(err.Error()).To(ContainSubstring("router short-circuit (test)"))
stampViaRouterCtx()
})
It("ImageGeneration forwards the request context to the SmartRouter", func() {
_, err := backend.ImageGeneration(reqCtx, 64, 64, 1, 0, "p", "", "", "/tmp/out.png", loader, modelCfg, appCfg, nil)
Expect(err).To(HaveOccurred())
Expect(err.Error()).To(ContainSubstring("router short-circuit (test)"))
stampViaRouterCtx()
})
It("does NOT leak the holder when the app context is used instead", func() {
// Sanity: the bug being fixed manifests as the router getting
// appCfg.Context (no holder) instead of reqCtx (holder). A direct
// call with context.Background() must not see the holder via the
// app context surface.
appCtxOnly := appCfg.Context
Expect(distributedhdr.Holder(appCtxOnly)).To(BeNil(),
"the app context must not be the carrier of per-request holders")
})
})