Add regexp matching of HTTP response headers to the http probe (#419)

Signed-off-by: Gleb Smirnov <[email protected]>

CONFIGURATION.md

@@ -58,14 +58,22 @@ The other placeholders are specified separately.
   # Probe fails if SSL is not present.
   [ fail_if_not_ssl: <boolean> | default = false ]
 
-  # Probe fails if response matches regex.
-  fail_if_matches_regexp:
+  # Probe fails if response body matches regex.
+  fail_if_body_matches_regexp:
     [ - <regex>, ... ]
 
-  # Probe fails if response does not match regex.
-  fail_if_not_matches_regexp:
+  # Probe fails if response body does not match regex.
+  fail_if_body_not_matches_regexp:
     [ - <regex>, ... ]
 
+  # Probe fails if response header matches regex. For headers with multiple values, fails if *at least one* matches.
+  fail_if_header_matches:
+    [ - <http_header_match_spec>, ... ]
+
+  # Probe fails if response header does not match regex. For headers with multiple values, fails if *none* match.
+  fail_if_header_not_matches:
+    [ - <http_header_match_spec>, ... ]
+
   # Configuration for TLS protocol of HTTP probe.
   tls_config:
     [ <tls_config> ]
@@ -86,14 +94,22 @@ The other placeholders are specified separately.
 
   # The IP protocol of the HTTP probe (ip4, ip6).
   [ preferred_ip_protocol: <string> | default = "ip6" ]
-  [ ip_protocol_fallback: <boolean | default = true> ]
+  [ ip_protocol_fallback: <boolean> | default = true ]
 
   # The body of the HTTP request used in probe.
   body: [ <string> ]
 
 
 ```
 
+#### <http_header_match_spec>
+
+```yml
+header: <string>,
+regexp: <regex>,
+[ allow_missing: <boolean> | default = false ]
+```
+
 ### <tcp_probe>
 
 ```yml

config/config.go

@@ -80,19 +80,27 @@ type Module struct {
 
 type HTTPProbe struct {
 	// Defaults to 2xx.
-	ValidStatusCodes       []int                   `yaml:"valid_status_codes,omitempty"`
-	ValidHTTPVersions      []string                `yaml:"valid_http_versions,omitempty"`
-	IPProtocol             string                  `yaml:"preferred_ip_protocol,omitempty"`
-	IPProtocolFallback     bool                    `yaml:"ip_protocol_fallback,omitempty"`
-	NoFollowRedirects      bool                    `yaml:"no_follow_redirects,omitempty"`
-	FailIfSSL              bool                    `yaml:"fail_if_ssl,omitempty"`
-	FailIfNotSSL           bool                    `yaml:"fail_if_not_ssl,omitempty"`
-	Method                 string                  `yaml:"method,omitempty"`
-	Headers                map[string]string       `yaml:"headers,omitempty"`
-	FailIfMatchesRegexp    []string                `yaml:"fail_if_matches_regexp,omitempty"`
-	FailIfNotMatchesRegexp []string                `yaml:"fail_if_not_matches_regexp,omitempty"`
-	Body                   string                  `yaml:"body,omitempty"`
-	HTTPClientConfig       config.HTTPClientConfig `yaml:"http_client_config,inline"`
+	ValidStatusCodes             []int                   `yaml:"valid_status_codes,omitempty"`
+	ValidHTTPVersions            []string                `yaml:"valid_http_versions,omitempty"`
+	IPProtocol                   string                  `yaml:"preferred_ip_protocol,omitempty"`
+	IPProtocolFallback           bool                    `yaml:"ip_protocol_fallback,omitempty"`
+	NoFollowRedirects            bool                    `yaml:"no_follow_redirects,omitempty"`
+	FailIfSSL                    bool                    `yaml:"fail_if_ssl,omitempty"`
+	FailIfNotSSL                 bool                    `yaml:"fail_if_not_ssl,omitempty"`
+	Method                       string                  `yaml:"method,omitempty"`
+	Headers                      map[string]string       `yaml:"headers,omitempty"`
+	FailIfBodyMatchesRegexp      []string                `yaml:"fail_if_body_matches_regexp,omitempty"`
+	FailIfBodyNotMatchesRegexp   []string                `yaml:"fail_if_body_not_matches_regexp,omitempty"`
+	FailIfHeaderMatchesRegexp    []HeaderMatch           `yaml:"fail_if_header_matches,omitempty"`
+	FailIfHeaderNotMatchesRegexp []HeaderMatch           `yaml:"fail_if_header_not_matches,omitempty"`
+	Body                         string                  `yaml:"body,omitempty"`
+	HTTPClientConfig             config.HTTPClientConfig `yaml:"http_client_config,inline"`
+}
+
+type HeaderMatch struct {
+	Header       string `yaml:"header,omitempty"`
+	Regexp       string `yaml:"regexp,omitempty"`
+	AllowMissing bool   `yaml:"allow_missing,omitempty"`
 }
 
 type QueryResponse struct {
@@ -217,3 +225,21 @@ func (s *QueryResponse) UnmarshalYAML(unmarshal func(interface{}) error) error {
 	}
 	return nil
 }
+
+// UnmarshalYAML implements the yaml.Unmarshaler interface.
+func (s *HeaderMatch) UnmarshalYAML(unmarshal func(interface{}) error) error {
+	type plain HeaderMatch
+	if err := unmarshal((*plain)(s)); err != nil {
+		return err
+	}
+
+	if s.Header == "" {
+		return errors.New("header name must be set for HTTP header matchers")
+	}
+
+	if s.Regexp == "" {
+		return errors.New("regexp must be set for HTTP header matchers")
+	}
+
+	return nil
+}

config/config_test.go

@@ -38,6 +38,10 @@ func TestLoadBadConfigs(t *testing.T) {
 			ConfigFile:    "testdata/invalid-dns-module.yml",
 			ExpectedError: "error parsing config file: query name must be set for DNS module",
 		},
+		{
+			ConfigFile:    "testdata/invalid-http-header-match.yml",
+			ExpectedError: "error parsing config file: regexp must be set for HTTP header matchers",
+		},
 	}
 	for i, test := range tests {
 		err := sc.ReloadConfig(test.ConfigFile)

config/testdata/blackbox-good.yml

@@ -66,3 +66,14 @@ modules:
       ip_protocol_fallback: false
       validate_answer_rrs:
         fail_if_matches_regexp: [test]
+  http_header_match_origin:
+    prober: http
+    timeout: 5s
+    http:
+      method: GET
+      headers:
+        Origin: example.com
+      fail_if_header_not_matches:
+        - header: Access-Control-Allow-Origin
+          regexp: '(\*|example\.com)'
+          allow_missing: false

config/testdata/invalid-http-header-match.yml

@@ -0,0 +1,8 @@
+modules:
+  http_headers:
+    prober: http
+    timeout: 5s
+    http:
+      fail_if_header_not_matches:
+        - header: Access-Control-Allow-Origin
+          allow_missing: false

example.yml

@@ -9,13 +9,21 @@ modules:
       headers:
         Host: vhost.example.com
         Accept-Language: en-US
+        Origin: example.com
       no_follow_redirects: false
       fail_if_ssl: false
       fail_if_not_ssl: false
-      fail_if_matches_regexp:
+      fail_if_body_matches_regexp:
         - "Could not connect to database"
-      fail_if_not_matches_regexp:
+      fail_if_body_not_matches_regexp:
         - "Download the latest version here"
+      fail_if_header_matches: # Verifies that no cookies are set
+        - header: Set-Cookie
+          allow_missing: true
+          regexp: '.*'
+      fail_if_header_not_matches:
+        - header: Access-Control-Allow-Origin
+          regexp: '(\*|example\.com)'
       tls_config:
         insecure_skip_verify: false
       preferred_ip_protocol: "ip4" # defaults to "ip6"

prober/http.go

@@ -23,6 +23,7 @@ import (
 	"net/http"
 	"net/http/cookiejar"
 	"net/http/httptrace"
+	"net/textproto"
 	"net/url"
 	"regexp"
 	"strconv"
@@ -44,7 +45,7 @@ func matchRegularExpressions(reader io.Reader, httpConfig config.HTTPProbe, logg
 		level.Error(logger).Log("msg", "Error reading HTTP body", "err", err)
 		return false
 	}
-	for _, expression := range httpConfig.FailIfMatchesRegexp {
+	for _, expression := range httpConfig.FailIfBodyMatchesRegexp {
 		re, err := regexp.Compile(expression)
 		if err != nil {
 			level.Error(logger).Log("msg", "Could not compile regular expression", "regexp", expression, "err", err)
@@ -55,7 +56,7 @@ func matchRegularExpressions(reader io.Reader, httpConfig config.HTTPProbe, logg
 			return false
 		}
 	}
-	for _, expression := range httpConfig.FailIfNotMatchesRegexp {
+	for _, expression := range httpConfig.FailIfBodyNotMatchesRegexp {
 		re, err := regexp.Compile(expression)
 		if err != nil {
 			level.Error(logger).Log("msg", "Could not compile regular expression", "regexp", expression, "err", err)
@@ -69,6 +70,68 @@ func matchRegularExpressions(reader io.Reader, httpConfig config.HTTPProbe, logg
 	return true
 }
 
+func matchRegularExpressionsOnHeaders(header http.Header, httpConfig config.HTTPProbe, logger log.Logger) bool {
+	for _, headerMatchSpec := range httpConfig.FailIfHeaderMatchesRegexp {
+		values := header[textproto.CanonicalMIMEHeaderKey(headerMatchSpec.Header)]
+		if len(values) == 0 {
+			if !headerMatchSpec.AllowMissing {
+				level.Error(logger).Log("msg", "Missing required header", "header", headerMatchSpec.Header)
+				return false
+			} else {
+				continue // No need to match any regex on missing headers.
+			}
+		}
+
+		re, err := regexp.Compile(headerMatchSpec.Regexp)
+		if err != nil {
+			level.Error(logger).Log("msg", "Could not compile regular expression", "regexp", headerMatchSpec.Regexp, "err", err)
+			return false
+		}
+
+		for _, val := range values {
+			if re.MatchString(val) {
+				level.Error(logger).Log("msg", "Header matched regular expression", "header", headerMatchSpec.Header,
+					"regexp", headerMatchSpec.Regexp, "value_count", len(values))
+				return false
+			}
+		}
+	}
+	for _, headerMatchSpec := range httpConfig.FailIfHeaderNotMatchesRegexp {
+		values := header[textproto.CanonicalMIMEHeaderKey(headerMatchSpec.Header)]
+		if len(values) == 0 {
+			if !headerMatchSpec.AllowMissing {
+				level.Error(logger).Log("msg", "Missing required header", "header", headerMatchSpec.Header)
+				return false
+			} else {
+				continue // No need to match any regex on missing headers.
+			}
+		}
+
+		re, err := regexp.Compile(headerMatchSpec.Regexp)
+		if err != nil {
+			level.Error(logger).Log("msg", "Could not compile regular expression", "regexp", headerMatchSpec.Regexp, "err", err)
+			return false
+		}
+
+		anyHeaderValueMatched := false
+
+		for _, val := range values {
+			if re.MatchString(val) {
+				anyHeaderValueMatched = true
+				break
+			}
+		}
+
+		if !anyHeaderValueMatched {
+			level.Error(logger).Log("msg", "Header did not match regular expression", "header", headerMatchSpec.Header,
+				"regexp", headerMatchSpec.Regexp, "value_count", len(values))
+			return false
+		}
+	}
+
+	return true
+}
+
 // roundTripTrace holds timings for a single HTTP roundtrip.
 type roundTripTrace struct {
 	tls           bool
@@ -320,7 +383,16 @@ func ProbeHTTP(ctx context.Context, target string, module config.Module, registr
 			level.Info(logger).Log("msg", "Invalid HTTP response status code, wanted 2xx", "status_code", resp.StatusCode)
 		}
 
-		if success && (len(httpConfig.FailIfMatchesRegexp) > 0 || len(httpConfig.FailIfNotMatchesRegexp) > 0) {
+		if success && (len(httpConfig.FailIfHeaderMatchesRegexp) > 0 || len(httpConfig.FailIfHeaderNotMatchesRegexp) > 0) {
+			success = matchRegularExpressionsOnHeaders(resp.Header, httpConfig, logger)
+			if success {
+				probeFailedDueToRegex.Set(0)
+			} else {
+				probeFailedDueToRegex.Set(1)
+			}
+		}
+
+		if success && (len(httpConfig.FailIfBodyMatchesRegexp) > 0 || len(httpConfig.FailIfBodyNotMatchesRegexp) > 0) {
 			success = matchRegularExpressions(resp.Body, httpConfig, logger)
 			if success {
 				probeFailedDueToRegex.Set(0)