x-pack/filebeat/input/cel: add default user-agent to http requests

CHANGELOG.next.asciidoc

@@ -267,6 +267,7 @@ https://github.com/elastic/beats/compare/v8.8.1\...main[Check the HEAD diff]
 - Update CEL mito extensions to v1.11.0 to improve type checking. {pull}39460[39460]
 - Improve logging of request and response with request trace logging in error conditions. {pull}39455[39455]
 - Add HTTP metrics to CEL input. {issue}39501[39501] {pull}39503[39503]
+- Add default user-agent to CEL HTTP requests. {issue}39502[39502] {pull}39587[39587]
 
 *Auditbeat*
 

x-pack/filebeat/docs/inputs/input-cel.asciidoc

@@ -242,7 +242,7 @@ As noted above the `cel` input provides functions, macros, and global variables
 * {mito_docs}/lib#Debug[Debug] — the debug handler registers a logger with the name extension `cel_debug` and calls to the CEL `debug` function are emitted to that logger.
 ** {mito_docs}/lib#hdr-Debug[Debug]
 
-In addition to the extensions provided in the packages listed above, a global variable `useragent` is also provided which gives the user CEL program access to the {beatname_lc} user-agent string.
+In addition to the extensions provided in the packages listed above, a global variable `useragent` is also provided which gives the user CEL program access to the {beatname_lc} user-agent string. By default, this value is assigned to all requests' user-agent headers unless the CEL program has already set the user-agent header value. Programs wishing to not provide a user-agent, should set this header to the empty string, `""`.
 
 The CEL environment enables the https://pkg.go.dev/github.com/google/cel-go/cel#OptionalTypes[optional types] library using the version defined {mito_docs}/lib#OptionalTypesVersion[here].
 

x-pack/filebeat/input/cel/input.go

@@ -64,7 +64,8 @@
 	root = "state"
 )
 
-// The Filebeat user-agent is provided to the program as useragent.
+// The Filebeat user-agent is provided to the program as useragent. If a request
+// is not given a user-agent string, this user agent is added to the request.
 var userAgent = useragent.UserAgent("Filebeat", version.GetDefaultVersion(), version.Commit(), version.BuildTime().String())
 
 func Plugin(log *logp.Logger, store inputcursor.StateStore) v2.Plugin {
@@ -758,6 +759,11 @@
 		return authClient, trace, nil
 	}
 
+	c.Transport = userAgentDecorator{
+		UserAgent: userAgent,
+		Transport: c.Transport,
+	}
+
 	return c, trace, nil
 }
 
@@ -856,6 +862,18 @@
 	}
 }
 
+type userAgentDecorator struct {
+	UserAgent string
+	Transport http.RoundTripper
+}
+
+func (t userAgentDecorator) RoundTrip(r *http.Request) (*http.Response, error) {
+	if _, ok := r.Header["User-Agent"]; !ok {
+		r.Header.Set("User-Agent", t.UserAgent)
+	}
+	return t.Transport.RoundTrip(r)
+}
+
 func newRateLimiterFromConfig(cfg *ResourceConfig) *rate.Limiter {
 	r := rate.Inf
 	b := 1
@@ -1166,7 +1184,7 @@
 // walkMap walks to all ends of the provided path in m and applies fn to the
 // final element of each walk. Nested arrays are not handled.
 func walkMap(m mapstr.M, path string, fn func(parent mapstr.M, key string)) {
 	key, rest, more := strings.Cut(path, ".")
 	v, ok := m[key]
 	if !ok {
 		return

x-pack/filebeat/input/cel/input_test.go

@@ -558,6 +558,142 @@ var inputTests = []struct {
 			},
 		},
 	},
+	{
+		name:   "GET_request_check_user_agent_default",
+		server: newTestServer(httptest.NewServer),
+		config: map[string]interface{}{
+			"interval": 1,
+			"program": `
+	get(state.url).Body.as(body, {
+		"events": [body.decode_json()]
+	})
+	`,
+		},
+		handler: func(w http.ResponseWriter, r *http.Request) {
+			w.Header().Set("content-type", "application/json")
+			msg := `{"hello":[{"world":"moon"},{"space":[{"cake":"pumpkin"}]}]}`
+			if r.UserAgent() != userAgent {
+				w.WriteHeader(http.StatusBadRequest)
+				msg = fmt.Sprintf(`{"error":"expected user agent was %#q"}`, userAgent)
+			}
+			if r.Method != http.MethodGet {
+				w.WriteHeader(http.StatusBadRequest)
+				msg = fmt.Sprintf(`{"error":"expected method was %#q"}`, http.MethodGet)
+			}
+
+			w.Write([]byte(msg))
+		},
+		want: []map[string]interface{}{
+			{
+				"hello": []interface{}{
+					map[string]interface{}{
+						"world": "moon",
+					},
+					map[string]interface{}{
+						"space": []interface{}{
+							map[string]interface{}{
+								"cake": "pumpkin",
+							},
+						},
+					},
+				},
+			},
+		},
+	},
+	{
+		name:   "GET_request_check_user_agent_user_defined",
+		server: newTestServer(httptest.NewServer),
+		config: map[string]interface{}{
+			"interval": 1,
+			"program": `
+	get_request(state.url).with({
+		"Header": {
+			"User-Agent": ["custom user agent"]
+		}
+	}).do_request().Body.as(body, {
+		"events": [body.decode_json()]
+	})
+	`,
+		},
+		handler: func(w http.ResponseWriter, r *http.Request) {
+			const customUserAgent = "custom user agent"
+
+			w.Header().Set("content-type", "application/json")
+			msg := `{"hello":[{"world":"moon"},{"space":[{"cake":"pumpkin"}]}]}`
+			if r.UserAgent() != customUserAgent {
+				w.WriteHeader(http.StatusBadRequest)
+				msg = fmt.Sprintf(`{"error":"expected user agent was %#q"}`, customUserAgent)
+			}
+			if r.Method != http.MethodGet {
+				w.WriteHeader(http.StatusBadRequest)
+				msg = fmt.Sprintf(`{"error":"expected method was %#q"}`, http.MethodGet)
+			}
+
+			w.Write([]byte(msg))
+		},
+		want: []map[string]interface{}{
+			{
+				"hello": []interface{}{
+					map[string]interface{}{
+						"world": "moon",
+					},
+					map[string]interface{}{
+						"space": []interface{}{
+							map[string]interface{}{
+								"cake": "pumpkin",
+							},
+						},
+					},
+				},
+			},
+		},
+	},
+	{
+		name:   "GET_request_check_user_agent_none",
+		server: newTestServer(httptest.NewServer),
+		config: map[string]interface{}{
+			"interval": 1,
+			"program": `
+	get_request(state.url).with({
+		"Header": {
+			"User-Agent": [""]
+		}
+	}).do_request().Body.as(body, {
+		"events": [body.decode_json()]
+	})
+	`,
+		},
+		handler: func(w http.ResponseWriter, r *http.Request) {
+			w.Header().Set("content-type", "application/json")
+			msg := `{"hello":[{"world":"moon"},{"space":[{"cake":"pumpkin"}]}]}`
+			if _, ok := r.Header["User-Agent"]; ok {
+				w.WriteHeader(http.StatusBadRequest)
+				msg = fmt.Sprintf(`{"error":"expected no user agent header, but got %#q"}`, r.UserAgent())
+			}
+			if r.Method != http.MethodGet {
+				w.WriteHeader(http.StatusBadRequest)
+				msg = fmt.Sprintf(`{"error":"expected method was %#q"}`, http.MethodGet)
+			}
+
+			w.Write([]byte(msg))
+		},
+		want: []map[string]interface{}{
+			{
+				"hello": []interface{}{
+					map[string]interface{}{
+						"world": "moon",
+					},
+					map[string]interface{}{
+						"space": []interface{}{
+							map[string]interface{}{
+								"cake": "pumpkin",
+							},
+						},
+					},
+				},
+			},
+		},
+	},
 	{
 		name:   "GET_request_TLS",
 		server: newTestServer(httptest.NewTLSServer),
@@ -1576,13 +1712,13 @@ func defaultHandler(expectedMethod, expectedBody string) http.HandlerFunc {
 		switch {
 		case r.Method != expectedMethod:
 			w.WriteHeader(http.StatusBadRequest)
-			msg = fmt.Sprintf(`{"error":"expected method was %q"}`, expectedMethod)
+			msg = fmt.Sprintf(`{"error":"expected method was %#q"}`, expectedMethod)
 		case expectedBody != "":
 			body, _ := io.ReadAll(r.Body)
 			r.Body.Close()
 			if expectedBody != string(body) {
 				w.WriteHeader(http.StatusBadRequest)
-				msg = fmt.Sprintf(`{"error":"expected body was %q"}`, expectedBody)
+				msg = fmt.Sprintf(`{"error":"expected body was %#q"}`, expectedBody)
 			}
 		}