Expose Envoy's /stats for statsd agents

command/connect/envoy/bootstrap_config.go

@@ -57,6 +57,14 @@ type BootstrapConfig struct {
 	// be fixed in a future Consul version as Envoy 1.10 reaches stable release.
 	PrometheusBindAddr string `mapstructure:"envoy_prometheus_bind_addr"`
 
+	// StatsBindAddr configures an <ip>:<port> on which the Envoy will listen
+	// and expose the /stats HTTP path prefix for any agent to access. It
+	// does this by proxying that path prefix to the internal admin server
+	// which allows exposing metrics on the network without the security
+	// risk of exposing the full admin server API. Any other URL requested will be
+	// a 404.
+	StatsBindAddr string `mapstructure:"envoy_stats_bind_addr"`
+
 	// OverrideJSONTpl allows replacing the base template used to render the
 	// bootstrap. This is an "escape hatch" allowing arbitrary control over the
 	// proxy's configuration but will the most effort to maintain and correctly
@@ -188,7 +196,13 @@ func (c *BootstrapConfig) ConfigureArgs(args *BootstrapTplArgs) error {
 	}
 	// Setup prometheus if needed. This MUST happen after the Static*JSON is set above
 	if c.PrometheusBindAddr != "" {
-		if err := c.generatePrometheusConfig(args); err != nil {
+		if err := c.generateMetricsListenerConfig(args, c.PrometheusBindAddr, "envoy_prometheus_metrics", "path", "/metrics", "/stats/prometheus"); err != nil {
+			return err
+		}
+	}
+	// Setup /stats proxy listener if needed. This MUST happen after the Static*JSON is set above
+	if c.StatsBindAddr != "" {
+		if err := c.generateMetricsListenerConfig(args, c.StatsBindAddr, "envoy_metrics", "prefix", "/stats", "/stats"); err != nil {
 			return err
 		}
 	}
@@ -369,10 +383,10 @@ func (c *BootstrapConfig) generateStatsConfig(args *BootstrapTplArgs) error {
 	return nil
 }
 
-func (c *BootstrapConfig) generatePrometheusConfig(args *BootstrapTplArgs) error {
-	host, port, err := net.SplitHostPort(c.PrometheusBindAddr)
+func (c *BootstrapConfig) generateMetricsListenerConfig(args *BootstrapTplArgs, bindAddr, name, matchType, matchValue, prefixRewrite string) error {
+	host, port, err := net.SplitHostPort(bindAddr)
 	if err != nil {
-		return fmt.Errorf("invalid prometheus_bind_addr: %s", err)
+		return fmt.Errorf("invalid %s bind address: %s", name, err)
 	}
 
 	clusterJSON := `{
@@ -390,7 +404,7 @@ func (c *BootstrapConfig) generatePrometheusConfig(args *BootstrapTplArgs) error
 		]
 	}`
 	listenerJSON := `{
-		"name": "envoy_prometheus_metrics_listener",
+		"name": "` + name + `_listener",
 		"address": {
 			"socket_address": {
 				"address": "` + host + `",
@@ -403,7 +417,7 @@ func (c *BootstrapConfig) generatePrometheusConfig(args *BootstrapTplArgs) error
 					{
 						"name": "envoy.http_connection_manager",
 						"config": {
-							"stat_prefix": "envoy_prometheus_metrics",
+							"stat_prefix": "` + name + `",
 							"codec_type": "HTTP1",
 							"route_config": {
 								"name": "self_admin_route",
@@ -416,11 +430,11 @@ func (c *BootstrapConfig) generatePrometheusConfig(args *BootstrapTplArgs) error
 										"routes": [
 											{
 												"match": {
-													"path": "/metrics"
+													"` + matchType + `": "` + matchValue + `"
 												},
 												"route": {
 													"cluster": "self_admin",
-													"prefix_rewrite": "/stats/prometheus"
+													"prefix_rewrite": "` + prefixRewrite + `"
 												}
 											},
 											{

command/connect/envoy/bootstrap_config_test.go

@@ -80,6 +80,77 @@ const (
 			}
 		]
 	}`
+	expectedStatsListener = `{
+		"name": "envoy_metrics_listener",
+		"address": {
+			"socket_address": {
+				"address": "0.0.0.0",
+				"port_value": 9000
+			}
+		},
+		"filter_chains": [
+			{
+				"filters": [
+					{
+						"name": "envoy.http_connection_manager",
+						"config": {
+							"stat_prefix": "envoy_metrics",
+							"codec_type": "HTTP1",
+							"route_config": {
+								"name": "self_admin_route",
+								"virtual_hosts": [
+									{
+										"name": "self_admin",
+										"domains": [
+											"*"
+										],
+										"routes": [
+											{
+												"match": {
+													"prefix": "/stats"
+												},
+												"route": {
+													"cluster": "self_admin",
+													"prefix_rewrite": "/stats"
+												}
+											},
+											{
+												"match": {
+													"prefix": "/"
+												},
+												"direct_response": {
+													"status": 404
+												}
+											}
+										]
+									}
+								]
+							},
+							"http_filters": [
+								{
+									"name": "envoy.router"
+								}
+							]
+						}
+					}
+				]
+			}
+		]
+	}`
+	expectedStatsCluster = `{
+		"name": "self_admin",
+		"connect_timeout": "5s",
+		"type": "STATIC",
+		"http_protocol_options": {},
+		"hosts": [
+			{
+				"socket_address": {
+					"address": "127.0.0.1",
+					"port_value": 19000
+				}
+			}
+		]
+	}`
 )
 
 func TestBootstrapConfig_ConfigureArgs(t *testing.T) {
@@ -350,6 +421,48 @@ func TestBootstrapConfig_ConfigureArgs(t *testing.T) {
 			},
 			wantErr: false,
 		},
+		{
+			name: "stats-bind-addr",
+			input: BootstrapConfig{
+				StatsBindAddr: "0.0.0.0:9000",
+			},
+			baseArgs: BootstrapTplArgs{
+				AdminBindAddress: "127.0.0.1",
+				AdminBindPort:    "19000",
+			},
+			wantArgs: BootstrapTplArgs{
+				AdminBindAddress: "127.0.0.1",
+				AdminBindPort:    "19000",
+				// Should add a static cluster for the self-proxy to admin
+				StaticClustersJSON: expectedStatsCluster,
+				// Should add a static http listener too
+				StaticListenersJSON: expectedStatsListener,
+				StatsConfigJSON:     defaultStatsConfigJSON,
+			},
+			wantErr: false,
+		},
+		{
+			name: "stats-bind-addr-with-overrides",
+			input: BootstrapConfig{
+				StatsBindAddr:       "0.0.0.0:9000",
+				StaticClustersJSON:  `{"foo":"bar"}`,
+				StaticListenersJSON: `{"baz":"qux"}`,
+			},
+			baseArgs: BootstrapTplArgs{
+				AdminBindAddress: "127.0.0.1",
+				AdminBindPort:    "19000",
+			},
+			wantArgs: BootstrapTplArgs{
+				AdminBindAddress: "127.0.0.1",
+				AdminBindPort:    "19000",
+				// Should add a static cluster for the self-proxy to admin
+				StaticClustersJSON: `{"foo":"bar"},` + expectedStatsCluster,
+				// Should add a static http listener too
+				StaticListenersJSON: `{"baz":"qux"},` + expectedStatsListener,
+				StatsConfigJSON:     defaultStatsConfigJSON,
+			},
+			wantErr: false,
+		},
 		{
 			name: "stats-flush-interval",
 			input: BootstrapConfig{
@@ -379,6 +492,13 @@ func TestBootstrapConfig_ConfigureArgs(t *testing.T) {
 			},
 			wantErr: true,
 		},
+		{
+			name: "err-bad-stats-addr",
+			input: BootstrapConfig{
+				StatsBindAddr: "asdasdsad",
+			},
+			wantErr: true,
+		},
 		{
 			name: "err-bad-statsd-addr",
 			input: BootstrapConfig{

test/integration/connect/envoy/case-stats-proxy/s1.hcl

@@ -0,0 +1,23 @@
+services {
+  name = "s1"
+  port = 8080
+  connect {
+    sidecar_service {
+      proxy {
+        upstreams = [
+          {
+            destination_name = "s2"
+            local_bind_port = 5000
+            config {
+              protocol = "http"
+            }
+          }
+        ]
+        config {
+          protocol = "http"
+          envoy_stats_bind_addr = "0.0.0.0:1239"
+        }
+      }
+    }
+  }
+}

test/integration/connect/envoy/case-stats-proxy/s2.hcl

@@ -0,0 +1,13 @@
+services {
+  name = "s2"
+  port = 8181
+  connect {
+    sidecar_service {
+      proxy {
+        config {
+          protocol = "http"
+        }
+      }
+    }
+  }
+}

test/integration/connect/envoy/case-stats-proxy/setup.sh

@@ -0,0 +1,6 @@
+#!/bin/bash
+
+set -eEuo pipefail
+
+gen_envoy_bootstrap s1 19000 primary
+gen_envoy_bootstrap s2 19001 primary

test/integration/connect/envoy/case-stats-proxy/verify.bats

@@ -0,0 +1,62 @@
+#!/usr/bin/env bats
+
+load helpers
+
+@test "s1 proxy admin is up on :19000" {
+  retry_default curl -f -s localhost:19000/stats -o /dev/null
+}
+
+@test "s2 proxy admin is up on :19001" {
+  retry_default curl -f -s localhost:19001/stats -o /dev/null
+}
+
+@test "s1 proxy listener should be up and have right cert" {
+  assert_proxy_presents_cert_uri localhost:21000 s1
+}
+
+@test "s2 proxy listener should be up and have right cert" {
+  assert_proxy_presents_cert_uri localhost:21001 s2
+}
+
+@test "s2 proxy should be healthy" {
+  assert_service_has_healthy_instances s2 1
+}
+
+@test "s1 upstream should have healthy endpoints for s2" {
+  # protocol is configured in an upstream override so the cluster name is customized here
+  assert_upstream_has_endpoints_in_status 127.0.0.1:19000 1a47f6e1~s2.default.primary HEALTHY 1
+}
+
+@test "s1 upstream should be able to connect to s2 with http/1.1" {
+  run retry_default curl --http1.1 -s -f -d hello localhost:5000
+  [ "$status" -eq 0 ]
+  [ "$output" = "hello" ]
+}
+
+@test "s1 proxy should be exposing the /stats prefix" {
+  # Should have http metrics. This is just a sample one. Require the metric to
+  # be present not just found in a comment (anchor the regexp).
+  retry_default \
+    must_match_in_stats_proxy_response localhost:1239 \
+    'stats' '^http.envoy_metrics.downstream_rq_active'
+
+  # Response should include the the local cluster request.
+  retry_default \
+    must_match_in_stats_proxy_response localhost:1239 \
+    'stats' 'cluster.local_agent.upstream_rq_active'
+
+  # Response should include the http public listener.
+  retry_default \
+     must_match_in_stats_proxy_response localhost:1239 \
+    'stats' 'http.public_listener_http'
+
+  # /stats/prometheus should also be reachable and labelling the local cluster.
+  retry_default \
+     must_match_in_stats_proxy_response localhost:1239 \
+    'stats/prometheus' '[\{,]local_cluster="s1"[,}]'
+
+  # /stats/prometheus should also be reachable and exposing metrics.
+  retry_default \
+     must_match_in_stats_proxy_response localhost:1239 \
+    'stats/prometheus' 'envoy_http_downstream_rq_active'
+}