fix(telemetry): record route patterns instead of request paths on HTTP metrics (#12557)

# Which Problems Are Solved

- The `uri` label of `http.server.request_count` and
`http.server.return_code_counter` holds the concrete request path, so
every object ID served over a REST path becomes its own metric series.
On our v4.16.2 instance `/debug/metrics` exposes 8,000 series, 4,467 of
which (56%) belong to `http_server_return_code_counter_total` alone,
spread over 4,390 distinct `uri` values such as
`/v2/sessions/385063742120926058`. The set only grows with the number of
objects the installation has served.
- Paths matching no route are recorded verbatim, so unauthenticated
scanner traffic (`/wp-json/`, `/login.php`, …) keeps adding series.
- The grpc-gateway answers `405 Method Not Allowed` without going
through the configured `errorHandler`, so those responses never reported
a route pattern, not even before the regression.

This is a regression, first shipped in v4.11.0; v4.10.1 is unaffected.
#9286 introduced the mechanism, #9523 extended it to unknown paths, and
#11435 removed it while reorganising the middleware packages:

> Removed setting of URI to context in metric middleware. There were
only setters and no getters. (Unused value)

The getter was `*recorder.RequestURI` in `RegisterRequestCounter` /
`RegisterRequestCodeCounter`, which the same PR replaced with
`baseURI(r)`. Both ends went at once, so nothing failed.

# How the Problems Are Solved

- `metrics.WithRequestURIPattern` / `metrics.SetRequestURIPattern` are
back, and `metrics.RequestURI` is now the single place deciding what to
label with: the pattern a router reported, or the requested path if none
did. `UnknownPath` moved to the metrics package so every surface shares
one constant.
- The HTTP metrics middleware prepares the context before passing the
request on, which is the half that got dropped.
- `setRequestURIPattern` in the grpc-gateway reports the pattern to
metrics again, not only to tracing, and is now also called on the `405`
branch.
- Requests routed by chi (the OIDC endpoints) fall back to the pattern
chi matched. That covers the RFC 7592 client configuration routes added
in #12315 (`/oauth/v2/register/{client_id}`), which are templated on the
client ID, and collapses unknown paths below the OIDC prefixes to
`UNKNOWN_PATH` as well.

Against the metrics dump from our instance,
`http_server_return_code_counter_total` drops from 4,467 series to
roughly 35, and the endpoint as a whole from 8,000 to about 3,570.

# Additional Changes

- Regression tests, which were missing and are the reason the revert
went unnoticed:
- `metrics`: resolution of the `uri` label, including that a pattern set
on a derived context reaches the middleware.
- `gateway`: every way the gateway can answer — success, error,
unroutable path, wrong method.
- `middleware`: both routing styles, i.e. a router reporting its own
pattern and a chi routed request.
- A doc comment on `SetRequestURIPattern` explaining that it is written
by routers and read back by the middleware through the context, so it
does not read as an unused setter again.

# Additional Context

- Closes #12556
- Restores #9286 and #9523
- Regressed in #11435

Co-authored-by: Cursor <[email protected]>

backend/v3/instrumentation/metrics/http_handler.go

@@ -1,6 +1,7 @@
 package metrics
 
 import (
+	"context"
 	"net/http"
 	"strings"
 
@@ -17,6 +18,11 @@ const (
 	Method                       = "method"
 	URI                          = "uri"
 	ReturnCode                   = "return_code"
+
+	// UnknownPath is recorded as uri for requests that did not match any route.
+	// Recording the requested path instead would allow anyone to create an unbounded
+	// number of metric series by calling arbitrary paths.
+	UnknownPath = "UNKNOWN_PATH"
 )
 
 type MetricType int32
@@ -31,9 +37,44 @@ type StatusRecorder interface {
 	Status() int
 }
 
+type requestURIKey struct{}
+
+// WithRequestURIPattern returns a context that carries the uri label of the HTTP metrics
+// recorded for the current request. The metrics middleware must call this before passing
+// the request on, so that the router serving it can substitute its route pattern through
+// [SetRequestURIPattern].
+func WithRequestURIPattern(ctx context.Context) context.Context {
+	var pattern string
+	return context.WithValue(ctx, requestURIKey{}, &pattern)
+}
+
+// SetRequestURIPattern records pattern as the uri label of the current request's HTTP
+// metrics, e.g. "/v2/sessions/{session_id}" instead of "/v2/sessions/385063742120926058".
+//
+// Routers that serve templated paths must call this for every response they produce,
+// including errors. Without it every resource ID gets its own metric series and the
+// number of series only grows with the number of objects in the installation.
+func SetRequestURIPattern(ctx context.Context, pattern string) {
+	uri, ok := ctx.Value(requestURIKey{}).(*string)
+	if !ok {
+		return
+	}
+	*uri = pattern
+}
+
+// RequestURI returns the uri label to record for r: the route pattern reported by the
+// router through [SetRequestURIPattern], or the requested path if the router did not
+// report one.
+func RequestURI(r *http.Request) string {
+	if pattern, ok := r.Context().Value(requestURIKey{}).(*string); ok && *pattern != "" {
+		return *pattern
+	}
+	return strings.Split(r.RequestURI, "?")[0]
+}
+
 func RegisterRequestCounter(recorder StatusRecorder, r *http.Request) {
 	var labels = map[string]attribute.Value{
-		URI:    attribute.StringValue(baseURI(r)),
+		URI:    attribute.StringValue(RequestURI(r)),
 		Method: attribute.StringValue(r.Method),
 	}
 	RegisterCounter(RequestCounter, RequestCountDescription)
@@ -47,14 +88,10 @@ func RegisterTotalRequestCounter(r *http.Request) {
 
 func RegisterRequestCodeCounter(recorder StatusRecorder, r *http.Request) {
 	var labels = map[string]attribute.Value{
-		URI:        attribute.StringValue(baseURI(r)),
+		URI:        attribute.StringValue(RequestURI(r)),
 		Method:     attribute.StringValue(r.Method),
 		ReturnCode: attribute.IntValue(recorder.Status()),
 	}
 	RegisterCounter(ReturnCodeCounter, ReturnCodeCounterDescription)
 	AddCount(r.Context(), ReturnCodeCounter, 1, labels)
 }
-
-func baseURI(r *http.Request) string {
-	return strings.Split(r.RequestURI, "?")[0]
-}

backend/v3/instrumentation/metrics/http_handler_test.go

@@ -0,0 +1,91 @@
+package metrics
+
+import (
+	"context"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestRequestURI(t *testing.T) {
+	tests := []struct {
+		name       string
+		requestURI string
+		// prepare returns the context the request is served with.
+		prepare func(context.Context) context.Context
+		want    string
+	}{
+		{
+			name:       "no pattern support, path is recorded",
+			requestURI: "/oauth/v2/token",
+			prepare:    func(ctx context.Context) context.Context { return ctx },
+			want:       "/oauth/v2/token",
+		},
+		{
+			name:       "no pattern support, query is stripped",
+			requestURI: "/oauth/v2/authorize?client_id=1234&state=abcd",
+			prepare:    func(ctx context.Context) context.Context { return ctx },
+			want:       "/oauth/v2/authorize",
+		},
+		{
+			name:       "pattern supported but not set, path is recorded",
+			requestURI: "/oauth/v2/token",
+			prepare:    WithRequestURIPattern,
+			want:       "/oauth/v2/token",
+		},
+		{
+			name:       "pattern set, pattern is recorded instead of the object id",
+			requestURI: "/v2/sessions/385063742120926058",
+			prepare: func(ctx context.Context) context.Context {
+				ctx = WithRequestURIPattern(ctx)
+				SetRequestURIPattern(ctx, "/v2/sessions/{session_id}")
+				return ctx
+			},
+			want: "/v2/sessions/{session_id}",
+		},
+		{
+			name:       "unknown path, no path is recorded",
+			requestURI: "/junk",
+			prepare: func(ctx context.Context) context.Context {
+				ctx = WithRequestURIPattern(ctx)
+				SetRequestURIPattern(ctx, UnknownPath)
+				return ctx
+			},
+			want: UnknownPath,
+		},
+		{
+			name:       "pattern set without context support, path is recorded",
+			requestURI: "/v2/sessions/385063742120926058",
+			prepare: func(ctx context.Context) context.Context {
+				SetRequestURIPattern(ctx, "/v2/sessions/{session_id}")
+				return ctx
+			},
+			want: "/v2/sessions/385063742120926058",
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			r := httptest.NewRequest(http.MethodGet, tt.requestURI, nil)
+			r = r.WithContext(tt.prepare(r.Context()))
+
+			assert.Equal(t, tt.want, RequestURI(r))
+		})
+	}
+}
+
+// TestSetRequestURIPattern_derivedContext ensures the pattern set by a handler is visible
+// to the middleware that seeded the context, even though the handler received a derived
+// context. Routers such as the grpc-gateway only ever report the pattern on their own
+// context, so a value copied by [context.WithValue] would not reach the caller.
+func TestSetRequestURIPattern_derivedContext(t *testing.T) {
+	r := httptest.NewRequest(http.MethodDelete, "/v2/sessions/385063742120926058", nil)
+	r = r.WithContext(WithRequestURIPattern(r.Context()))
+
+	type unrelatedKey struct{}
+	derived := context.WithValue(r.Context(), unrelatedKey{}, "value")
+	SetRequestURIPattern(derived, "/v2/sessions/{session_id}")
+
+	assert.Equal(t, "/v2/sessions/{session_id}", RequestURI(r))
+}