From 9160789b421ae1d28c813c8ed3297cedd469722a Mon Sep 17 00:00:00 2001
From: Matthew Holt <mholt@users.noreply.github.com>
Date: Thu, 10 May 2018 08:57:25 -0600
Subject: [PATCH] telemetry: Make http_user_agent a normalized field
This way we store a short 8-byte hash of the UA instead of the full
string; exactly the same way we store TLS ClientHello info.
---
caddyhttp/httpserver/mitm.go | 3 ++-
caddyhttp/httpserver/server.go | 4 +++-
caddytls/handshake.go | 2 +-
telemetry/collection.go | 11 +++++++++++
4 files changed, 17 insertions(+), 3 deletions(-)
diff --git a/caddyhttp/httpserver/mitm.go b/caddyhttp/httpserver/mitm.go
index 6744a924e..d2faf5f3f 100644
--- a/caddyhttp/httpserver/mitm.go
+++ b/caddyhttp/httpserver/mitm.go
@@ -65,9 +65,10 @@ func (h *tlsHandler) ServeHTTP(w http.ResponseWriter, r *http.Request) {
h.listener.helloInfosMu.RUnlock()
ua := r.Header.Get("User-Agent")
+ uaHash := telemetry.FastHash([]byte(ua))
// report this request's UA in connection with this ClientHello
- go telemetry.AppendUnique("tls_client_hello_ua:"+caddytls.ClientHelloInfo(info).Key(), ua)
+ go telemetry.AppendUnique("tls_client_hello_ua:"+caddytls.ClientHelloInfo(info).Key(), uaHash)
var checked, mitm bool
if r.Header.Get("X-BlueCoat-Via") != "" || // Blue Coat (masks User-Agent header to generic values)
diff --git a/caddyhttp/httpserver/server.go b/caddyhttp/httpserver/server.go
index 45f1c639d..800f921de 100644
--- a/caddyhttp/httpserver/server.go
+++ b/caddyhttp/httpserver/server.go
@@ -354,7 +354,9 @@ func (s *Server) ServeHTTP(w http.ResponseWriter, r *http.Request) {
if len(ua) > 512 {
ua = ua[:512]
}
- go telemetry.AppendUnique("http_user_agent", ua)
+ uaHash := telemetry.FastHash([]byte(ua)) // this is a normalized field
+ go telemetry.SetNested("http_user_agent", uaHash, ua)
+ go telemetry.AppendUnique("http_user_agent_count", uaHash)
go telemetry.Increment("http_request_count")
// copy the original, unchanged URL into the context
diff --git a/caddytls/handshake.go b/caddytls/handshake.go
index 8b2639845..7e2ccb95e 100644
--- a/caddytls/handshake.go
+++ b/caddytls/handshake.go
@@ -516,7 +516,7 @@ func (info ClientHelloInfo) Key() string {
if !info.CompressionMethodsUnknown {
compressionMethods = fmt.Sprintf("%x", info.CompressionMethods)
}
- return fastHash([]byte(fmt.Sprintf("%x-%x-%s-%s-%x-%x",
+ return telemetry.FastHash([]byte(fmt.Sprintf("%x-%x-%s-%s-%x-%x",
info.Version, info.CipherSuites, extensions,
compressionMethods, info.Curves, info.Points)))
}
diff --git a/telemetry/collection.go b/telemetry/collection.go
index a46e2caf1..07cf0dc9c 100644
--- a/telemetry/collection.go
+++ b/telemetry/collection.go
@@ -15,6 +15,8 @@
package telemetry
import (
+ "fmt"
+ "hash/fnv"
"log"
"strings"
@@ -276,6 +278,15 @@ func atomicAdd(key string, amount int) {
bufferMu.Unlock()
}
+// FastHash hashes input using a 32-bit hashing algorithm
+// that is fast, and returns the hash as a hex-encoded string.
+// Do not use this for cryptographic purposes.
+func FastHash(input []byte) string {
+ h := fnv.New32a()
+ h.Write(input)
+ return fmt.Sprintf("%x", h.Sum32())
+}
+
// isDisabled returns whether key is
// a disabled metric key. ALL collection
// functions should call this and not