From d08efc34c38a951ff02c42dd007b77e94c809e01 Mon Sep 17 00:00:00 2001
From: Cecylia Bocovich <cohosh@torproject.org>
Date: Tue, 9 Sep 2025 12:41:27 -0400
Subject: [PATCH] Add prometheus metric for proxy answer counts

This adds a prometheus metric that tracks snowflake proxy answers. If
the client has not timed out before the proxy responds with an answer,
the proxy type is recorded along with a status of "success". If the
client has timed out, the type is left blank and the status is recorded
as "timeout".

The goal of these metrics is to help us determine how many proxies fail
to respond and to help narrow down which proxy implementations are
causing client timeouts.
---
 broker/ipc.go     |  2 ++
 broker/metrics.go | 12 +++++++++++-
 2 files changed, 13 insertions(+), 1 deletion(-)

diff --git a/broker/ipc.go b/broker/ipc.go
index 728363a..7e2951c 100644
--- a/broker/ipc.go
+++ b/broker/ipc.go
@@ -253,6 +253,7 @@ func (i *IPC) ProxyAnswers(arg messages.Arg, response *[]byte) error {
 		// The snowflake took too long to respond with an answer, so its client
 		// disappeared / the snowflake is no longer recognized by the Broker.
 		success = false
+		i.ctx.metrics.promMetrics.ProxyAnswerTotal.With(prometheus.Labels{"type": "", "status": "timeout"}).Inc()
 	}
 
 	b, err := messages.EncodeAnswerResponse(success)
@@ -263,6 +264,7 @@ func (i *IPC) ProxyAnswers(arg messages.Arg, response *[]byte) error {
 	*response = b
 
 	if success {
+		i.ctx.metrics.promMetrics.ProxyAnswerTotal.With(prometheus.Labels{"type": snowflake.proxyType, "status": "success"}).Inc()
 		snowflake.answerChannel <- answer
 	}
 
diff --git a/broker/metrics.go b/broker/metrics.go
index a09691b..ae1183c 100644
--- a/broker/metrics.go
+++ b/broker/metrics.go
@@ -290,6 +290,7 @@ type PromMetrics struct {
 	ProxyTotal       *prometheus.CounterVec
 	ProxyPollTotal   *safeprom.CounterVec
 	ClientPollTotal  *safeprom.CounterVec
+	ProxyAnswerTotal *safeprom.CounterVec
 	AvailableProxies *prometheus.GaugeVec
 
 	ProxyPollWithRelayURLExtensionTotal    *safeprom.CounterVec
@@ -331,6 +332,15 @@ func initPrometheus() *PromMetrics {
 		[]string{"nat", "status"},
 	)
 
+	promMetrics.ProxyAnswerTotal = safeprom.NewCounterVec(
+		prometheus.CounterOpts{
+			Namespace: prometheusNamespace,
+			Name:      "rounded_proxy_answer_total",
+			Help:      "The number of snowflake proxy answers, rounded up to a multiple of 8",
+		},
+		[]string{"type", "status"},
+	)
+
 	promMetrics.ProxyPollWithRelayURLExtensionTotal = safeprom.NewCounterVec(
 		prometheus.CounterOpts{
 			Namespace: prometheusNamespace,
@@ -370,7 +380,7 @@ func initPrometheus() *PromMetrics {
 	// We need to register our metrics so they can be exported.
 	promMetrics.registry.MustRegister(
 		promMetrics.ClientPollTotal, promMetrics.ProxyPollTotal,
-		promMetrics.ProxyTotal, promMetrics.AvailableProxies,
+		promMetrics.ProxyTotal, promMetrics.ProxyAnswerTotal, promMetrics.AvailableProxies,
 		promMetrics.ProxyPollWithRelayURLExtensionTotal,
 		promMetrics.ProxyPollWithoutRelayURLExtensionTotal,
 		promMetrics.ProxyPollRejectedForRelayURLExtensionTotal,