Metrics are served via metrics.status.sslip.io

- The metrics aren't fleshed out. In fact, there's only two so far:
  1. uptime
  2. number of queries
- Even though the metrics aren't complete, I'm checking it in because
  this commit is already much too big.
- I moved the version information to `version.status.sslip.io`;
  previously it was at `version.sslip.io`. I didn't want one endpoint
  for both metrics & version (worry: DNS amplification), and I wanted a
  consistent subdomain to find that information (i.e.
  `status.sslip.io`).
- I'm not worried about atomic updates to the metrics; if a metric is
  off by one, if I skip a count because two lookups are happening at the
  exact same time, I don't care.
- The `Metrics` struct is a pointer within `Xip` because I might have
  several copies of `Xip` (if I'm binding to several interfaces
  individually), but I must only have one copy of `Metrics`
- I only include the metrics I'm interested in, usually because it took
  some work to implement that feature. I don't care about MX records,
  but I care about IPv6 lookups, DNS-01 challenges, public IP lookups.
- got rid of a section of unreachable code at the end of
  `ProcessQuestion()`; I was tired of Goland flagging it. I had it there
  mostly because I was paranoid of falling through a `switch` statement

bosh-release/src/sslip.io-dns-server/go.mod

@@ -25,7 +25,7 @@ require (
 	golang.org/x/text v0.3.7 // indirect
 	golang.org/x/tools v0.1.8 // indirect
 	golang.org/x/xerrors v0.0.0-20200804184101-5ec99f83aff1 // indirect
-	google.golang.org/genproto v0.0.0-20220114231437-d2e6a121cae0 // indirect
+	google.golang.org/genproto v0.0.0-20220118154757-00ab72f36ad5 // indirect
 	google.golang.org/grpc v1.43.0 // indirect
 	google.golang.org/protobuf v1.27.1 // indirect
 	gopkg.in/yaml.v2 v2.4.0 // indirect

bosh-release/src/sslip.io-dns-server/go.sum

@@ -301,8 +301,8 @@ google.golang.org/genproto v0.0.0-20190819201941-24fa4b261c55/go.mod h1:DMBHOl98
 google.golang.org/genproto v0.0.0-20200513103714-09dca8ec2884/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
 google.golang.org/genproto v0.0.0-20200526211855-cb27e3aa2013/go.mod h1:NbSheEEYHJ7i3ixzK3sjbqSGDJWnxyFXZblF3eUsNvo=
 google.golang.org/genproto v0.0.0-20210602131652-f16073e35f0c/go.mod h1:UODoCrxHCcBojKKwX1terBiRUaqAsFqJiF615XL43r0=
-google.golang.org/genproto v0.0.0-20220114231437-d2e6a121cae0 h1:aCsSLXylHWFno0r4S3joLpiaWayvqd2Mn4iSvx4WZZc=
+google.golang.org/genproto v0.0.0-20220118154757-00ab72f36ad5 h1:zzNejm+EgrbLfDZ6lu9Uud2IVvHySPl8vQzf04laR5Q=
-google.golang.org/genproto v0.0.0-20220114231437-d2e6a121cae0/go.mod h1:5CzLGKJ67TSI2B9POpiiyGha0AjJvZIUgRMt1dSmuhc=
+google.golang.org/genproto v0.0.0-20220118154757-00ab72f36ad5/go.mod h1:5CzLGKJ67TSI2B9POpiiyGha0AjJvZIUgRMt1dSmuhc=
 google.golang.org/grpc v1.19.0/go.mod h1:mqu4LbDTu4XGKhr4mRzUsmM4RtVoemTSY81AxZiDr8c=
 google.golang.org/grpc v1.23.0/go.mod h1:Y5yQAOtifL1yxbo5wqy6BxZv8vAUGQwXBOALyacEbxg=
 google.golang.org/grpc v1.25.1/go.mod h1:c3i+UQWmh7LiEpx4sFZnkU36qjEYZ0imhYfXVyQciAY=

bosh-release/src/sslip.io-dns-server/integration_metrics_test.go

@@ -0,0 +1,58 @@
+package main_test
+
+import (
+	"fmt"
+	"os/exec"
+	"strings"
+	"time"
+	"xip/xip"
+
+	. "github.com/onsi/ginkgo/v2"
+
+	. "github.com/onsi/gomega"
+	. "github.com/onsi/gomega/gexec"
+)
+
+var _ = Describe("IntegrationMetrics", func() {
+	var digCmd *exec.Cmd
+	var digSession *Session
+	var digArgs string
+
+	When("the server is queried", func() {
+		It("should update metrics", func() {
+			startMetrics := getMetrics()
+			digArgs = "@localhost non-existent.sslip.io +short"
+			digCmd = exec.Command("dig", strings.Split(digArgs, " ")...)
+			digSession, err = Start(digCmd, GinkgoWriter, GinkgoWriter)
+			Expect(err).ToNot(HaveOccurred())
+			// we want to make sure digSession has exited because we
+			// want to parse the _full_ contents of stdout
+			Eventually(digSession, 1).Should(Exit(0))
+			expectedMetrics := startMetrics
+			expectedMetrics.Queries += 2 // two queries: nonexistent.sslip.io, metrics.status.sslip.io
+			actualMetrics := getMetrics()
+			Expect(expectedMetrics.MostlyEquals(actualMetrics)).To(BeTrue())
+		})
+	})
+})
+
+func getMetrics() (m xip.Metrics) {
+	digArgs := "@localhost metrics.status.sslip.io txt +short"
+	digCmd := exec.Command("dig", strings.Split(digArgs, " ")...)
+	stdout, err := digCmd.Output()
+	Expect(err).ToNot(HaveOccurred())
+	var uptime int
+	var junk string
+	_, err = fmt.Sscanf(string(stdout),
+		"\"uptime (seconds): %d\"\n"+
+			"\"key-value store: %s\n"+ // %s "swallows" the double-quote at the end
+			"\"queries: %d\"\n",
+		&uptime,
+		&junk,
+		&m.Queries,
+	)
+	Expect(err).ToNot(HaveOccurred())
+	m.Start = time.Now().Add(-time.Duration(uptime) * time.Second)
+	//_, err = fmt.Fscanf(digSession.Out, "queries: %d", &m.Queries)
+	return m
+}

bosh-release/src/sslip.io-dns-server/integration_test.go

@@ -109,10 +109,10 @@ var _ = Describe("sslip.io-dns-server", func() {
 				"@localhost example.com srv +short",
 				`\A\z`,
 				`TypeSRV example.com. \? nil, SOA example.com. briancunnie.gmail.com. 2021123100 900 900 1800 300\n$`),
-			Entry(`TXT for version.sslip.io is the version number of the xip software (which gets overwritten during linking)`,
+			Entry(`TXT for version.status.sslip.io is the version number of the xip software (which gets overwritten during linking)`,
-				"@127.0.0.1 version.sslip.io txt +short",
+				"@127.0.0.1 version.status.sslip.io txt +short",
 				`\A"dev"\n"today"\n"xxx"\n\z`,
-				`TypeTXT version.sslip.io. \? \["dev"\], \["today"\], \["xxx"\]`),
+				`TypeTXT version.status.sslip.io. \? \["dev"\], \["today"\], \["xxx"\]`),
 			Entry(`TXT is the querier's IPv4 address and the domain "ip.sslip.io"`,
 				"@127.0.0.1 ip.sslip.io txt +short",
 				`127.0.0.1`,

bosh-release/src/sslip.io-dns-server/main.go

@@ -33,12 +33,14 @@ func main() {
 	}
 	// I don't need to `defer etcdCli.Close()` it's redundant in the main routine: when main() exits, everything is closed.
 	conn, err := net.ListenUDP("udp", &net.UDPAddr{Port: 53})
+	// set up our global metrics struct, setting our start time
+	xipMetrics := xip.Metrics{Start: time.Now()}
 	//  common err hierarchy: net.OpError → os.SyscallError → syscall.Errno
 	switch {
 	case err == nil:
 		log.Println(`Successfully bound to all interfaces, port 53.`)
 		wg.Add(1)
-		readFrom(conn, etcdCli, &wg)
+		readFrom(conn, &wg, etcdCli, &xipMetrics)
 	case isErrorPermissionsError(err):
 		log.Println("Try invoking me with `sudo` because I don't have permission to bind to port 53.")
 		log.Fatal(err.Error())
@@ -62,7 +64,7 @@ func main() {
 			} else {
 				wg.Add(1)
 				boundIPsPorts = append(boundIPsPorts, conn.LocalAddr().String())
-				go readFrom(conn, etcdCli, &wg)
+				go readFrom(conn, &wg, etcdCli, &xipMetrics)
 			}
 		}
 		if len(boundIPsPorts) > 0 {
@@ -77,7 +79,7 @@ func main() {
 	wg.Wait()
 }
 
-func readFrom(conn *net.UDPConn, etcdCli *clientv3.Client, wg *sync.WaitGroup) {
+func readFrom(conn *net.UDPConn, wg *sync.WaitGroup, etcdCli xip.V3client, xipMetrics *xip.Metrics) {
 	defer wg.Done()
 	for {
 		query := make([]byte, 512)
@@ -87,7 +89,8 @@ func readFrom(conn *net.UDPConn, etcdCli *clientv3.Client, wg *sync.WaitGroup) {
 			continue
 		}
 		go func() {
-			response, logMessage, err := xip.Xip{SrcAddr: addr.IP, Etcd: etcdCli}.QueryResponse(query)
+			xipServer := xip.Xip{SrcAddr: addr.IP, Etcd: etcdCli, Metrics: xipMetrics}
+			response, logMessage, err := xipServer.QueryResponse(query)
 			if err != nil {
 				log.Println(err.Error())
 				return

bosh-release/src/sslip.io-dns-server/xip/xip.go

@@ -8,6 +8,7 @@ import (
 	"errors"
 	"fmt"
 	"net"
+	"reflect"
 	"regexp"
 	"strconv"
 	"strings"
@@ -33,6 +34,18 @@ type V3client interface {
 type Xip struct {
 	SrcAddr net.IP
 	Etcd    V3client
+	Metrics *Metrics
+}
+
+type Metrics struct {
+	Start                              time.Time
+	Queries                            int
+	SuccessfulQueries                  int
+	SuccessfulAQueries                 int
+	SuccessfulAAAAQueries              int
+	SuccessfulTXTSrcIPQueries          int
+	SuccessfulTXTVersionQueries        int
+	SuccessfulTXTDNS01ChallengeQUeries int
 }
 
 // DomainCustomization is a value that is returned for a specific query.
@@ -48,7 +61,7 @@ type DomainCustomization struct {
 	AAAA  []dnsmessage.AAAAResource
 	CNAME dnsmessage.CNAMEResource
 	MX    []dnsmessage.MXResource
-	TXT   func(string) ([]dnsmessage.TXTResource, error)
+	TXT   func(Xip) ([]dnsmessage.TXTResource, error)
 	// Unlike the other record types, TXT is a function in order to enable more complex behavior
 	// e.g. IP address of the query's source
 }
@@ -115,7 +128,7 @@ var (
 					MX:   mx2,
 				},
 			},
-			TXT: func(_ string) ([]dnsmessage.TXTResource, error) {
+			TXT: func(_ Xip) ([]dnsmessage.TXTResource, error) {
 				// Although multiple TXT records with multiple strings are allowed, we're sticking
 				// with a multiple TXT records with a single string apiece because that's what ProtonMail requires
 				// and that's what google.com does.
@@ -161,8 +174,8 @@ var (
 		"ip.sslip.io.": {
 			TXT: ipSslipIo,
 		},
-		"version.sslip.io.": {
+		"version.status.sslip.io.": {
-			TXT: func(_ string) ([]dnsmessage.TXTResource, error) {
+			TXT: func(_ Xip) ([]dnsmessage.TXTResource, error) {
 				return []dnsmessage.TXTResource{
 					{TXT: []string{VersionSemantic}}, // e.g. "2.2.1'
 					{TXT: []string{VersionDate}},     // e.g. "2021/10/03-15:08:54+0100"
@@ -170,6 +183,9 @@ var (
 				}, nil
 			},
 		},
+		"metrics.status.sslip.io.": {
+			TXT: metricsSslipIo,
+		},
 	}
 )
 
@@ -220,6 +236,7 @@ func (x Xip) QueryResponse(queryBytes []byte) (responseBytes []byte, logMessage
 	response.Header.ID = queryHeader.ID
 	response.Header.RecursionDesired = queryHeader.RecursionDesired
 
+	x.Metrics.Queries += 1
 	b := dnsmessage.NewBuilder(nil, response.Header)
 	b.EnableCompression()
 	if err = b.StartQuestions(); err != nil {
@@ -538,8 +555,6 @@ func (x Xip) processQuestion(q dnsmessage.Question) (response Response, logMessa
 			return response, logMessage + "nil, SOA " + soaLogMessage(soaResource), nil
 		}
 	}
-	// The following is flagged as "Unreachable code" in Goland, and that's expected
-	return response, "", errors.New("unexpectedly fell through x.processQuestion()")
 }
 
 // NSResponse sets the Answers/Authorities depending whether we're delegating or authoritative
@@ -720,7 +735,7 @@ func (x Xip) TXTResources(fqdn string) ([]dnsmessage.TXTResource, error) {
 	if domain, ok := Customizations[strings.ToLower(fqdn)]; ok {
 		// Customizations[strings.ToLower(fqdn)] returns a _function_,
 		// we call that function, which has the same return signature as this method
-		return domain.TXT(x.SrcAddr.String())
+		return domain.TXT(x)
 	}
 	return nil, nil
 }
@@ -750,8 +765,28 @@ func SOAResource(name dnsmessage.Name) dnsmessage.SOAResource {
 }
 
 // when TXT for "ip.sslip.io" is queried, return the IP address of the querier
-func ipSslipIo(sourceIP string) ([]dnsmessage.TXTResource, error) {
+func ipSslipIo(x Xip) ([]dnsmessage.TXTResource, error) {
-	return []dnsmessage.TXTResource{{TXT: []string{sourceIP}}}, nil
+	return []dnsmessage.TXTResource{{TXT: []string{x.SrcAddr.String()}}}, nil
+}
+
+// when TXT for "metrics.sslip.io" is queried, return the cumulative metrics
+func metricsSslipIo(x Xip) (txtResources []dnsmessage.TXTResource, err error) {
+	var metrics []string
+	uptime := time.Since(x.Metrics.Start)
+	metrics = append(metrics, fmt.Sprintf("uptime (seconds): %.0f", uptime.Seconds()))
+	keyValueStore := "etcd"
+	// comparing interfaces to nil are tricky: interfaces contain both a type
+	// and a value, and although the value is nil the type isn't, so we need the following
+	if x.Etcd == nil || reflect.ValueOf(x.Etcd).IsNil() {
+		keyValueStore = "builtin"
+	}
+	metrics = append(metrics, "key-value store: "+keyValueStore)
+	metrics = append(metrics, fmt.Sprintf("queries: %d", x.Metrics.Queries))
+	metrics = append(metrics, fmt.Sprintf("queries/second: %.1f", float64(x.Metrics.Queries)/uptime.Seconds()))
+	for _, metric := range metrics {
+		txtResources = append(txtResources, dnsmessage.TXTResource{TXT: []string{metric}})
+	}
+	return txtResources, nil
 }
 
 // when TXT for "k-v.io" is queried, return the key-value pair
@@ -868,3 +903,17 @@ func soaLogMessage(soaResource dnsmessage.SOAResource) string {
 		strconv.Itoa(int(soaResource.Expire)) + " " +
 		strconv.Itoa(int(soaResource.MinTTL))
 }
+
+// MostlyEquals compares all fields except `Start` (timestamp)
+func (a Metrics) MostlyEquals(b Metrics) bool {
+	if a.Queries == b.Queries &&
+		a.SuccessfulQueries == b.SuccessfulQueries &&
+		a.SuccessfulAQueries == b.SuccessfulAQueries &&
+		a.SuccessfulAAAAQueries == b.SuccessfulAAAAQueries &&
+		a.SuccessfulTXTSrcIPQueries == b.SuccessfulTXTSrcIPQueries &&
+		a.SuccessfulTXTVersionQueries == b.SuccessfulTXTVersionQueries &&
+		a.SuccessfulTXTDNS01ChallengeQUeries == b.SuccessfulTXTDNS01ChallengeQUeries {
+		return true
+	}
+	return false
+}

docs/DEVELOPER.md

@@ -8,7 +8,7 @@ export OLD_VERSION=2.2.4
 export VERSION=2.3.0
 cd ~/workspace/sslip.io
 git pull -r --autostash
-# update the version number for the TXT record for version.sslip.io
+# update the version number for the TXT record for version.status.sslip.io
 sed -i '' "s/$OLD_VERSION/$VERSION/g" \
   bin/make_all \
   bosh-release/packages/sslip.io-dns-server/packaging \
@@ -17,7 +17,7 @@ sed -i '' "s/$OLD_VERSION/$VERSION/g" \
 sed -i '' "s~/$OLD_VERSION/~/$VERSION/~g" \
   k8s/document_root/index.html \
   k8s/Dockerfile-sslip.io-dns-server
-# update the git hash for the TXT record for version.sslip.io for BOSH release
+# update the git hash for the TXT record for version.status.sslip.io for BOSH release
 sed -i '' "s/VersionGitHash=[0-9a-fA-F]*/VersionGitHash=$(git rev-parse --short HEAD)/g" \
   bosh-release/packages/sslip.io-dns-server/packaging
 cd bosh-release/
@@ -49,7 +49,7 @@ dig +short sSlIp.Io
 echo 78.46.204.247
 dig @$IP txt ip.sslip.io +short | tr -d '"'
 curl curlmyip.org; echo
-dig @$IP txt version.sslip.io +short | grep $VERSION
+dig @$IP txt version.status.sslip.io +short | grep $VERSION
 echo "\"$VERSION\""
 dig @$IP my-key.kv.sslip.io txt +short # returns nothing
 echo " ===" # separator because the results are too similar
@@ -101,7 +101,7 @@ git pull -r
 nvim sslip.io.yml
 bosh -e vsphere -d sslip.io deploy sslip.io.yml -l <(lpass show --note deployments.yml) --no-redact
 dig @ns-azure 127-0-0-1.sslip.io +short  # output should be 127.0.0.1
-dig @ns-azure.nono.io txt version.sslip.io +short
+dig @ns-azure.nono.io txt version.status.sslip.io +short
 git add -p
 git ci -v -m"Bump sslip.io BOSH release: $OLD_VERSION → $VERSION"
 git push

k8s/document_root/index.html

@@ -226,7 +226,7 @@ dig @ns.sslip.io txt ip.sslip.io +short -6 # forces IPv6 lookup; sample reply "2
       <code>ns-aws.sslip.io</code> requires a mere 592 bytes spread over 2 packets; Querying <a href=
       "https://icanhazip.com/">https://icanhazip.com/</a> requires 8692 bytes spread out over 34 packets—over 14 times
       as much! Admittedly bandwidth usage is a bigger concern for the one hosting the service than the one using the
-      service.</p>
+      service.</p><!--
       <h4 id="key-value-store"><code>k-v.io</code>: (key-value) read/write/delete TXTs</h4>
       <p>We enable special behavior under the <code>k-v.io</code> domain: it can be treated as a key-value store, the
       subdomain being the key, and the TXT record being the value.</p>
@@ -260,10 +260,11 @@ dig @ns.sslip.io txt ip.sslip.io +short -6 # forces IPv6 lookup; sample reply "2
         <li>Values are not consistent. If a value is set in <code>ns-aws.sslip.io</code>, it does not propagate to
         <code>ns-gce.sslip.io</code> nor <code>ns-azure.sslip.io</code>.</li>
       </ul>
+      -->
       <h4 id="version">Determining The Server Version of Software</h4>You can determine the server version of the
-      sslip.io software by querying the TXT record of <code>version.sslip.io</code>:
+      sslip.io software by querying the TXT record of <code>version.status.sslip.io</code>:
       <pre>
-dig @ns-aws.sslip.io txt version.sslip.io +short
+dig @ns-aws.sslip.io txt version.status.sslip.io +short
             "2.2.3"
             "2021/11/27-11:35:50-0800"
             "074f0a8"
@@ -271,10 +272,28 @@ dig @ns-aws.sslip.io txt version.sslip.io +short
       <p>The first number, ("2.2.3"), is the version of the sslip.io DNS software, and is most relevant. The other two
       numbers are the date compiled and the most recent git hash, but those values can differ across servers due to the
       manner in which the software is deployed.</p>
+      <h4 id="metrics">Server Metrics</h4>You can retrieve metrics from a given server by querying the TXT records of
+      <code>metrics.status.sslip.io</code>
+      <pre>
+dig @ns-azure.sslip.io txt version.status.sslip.io +short
+            "uptime (seconds): 1200"
+            "key-value store: builtin"
+            "queries: 46202"
+            "queries/second: 38.5"
+            "successful:"
+            "- queries/second: 14.5"
+            "- A: 2000"
+            "- AAAA: 20"
+            "- IP TXT: 2"
+            "- version TXT: 2"
+            "- DNS-01 challenge: 2"
+      </pre>
       <h3 id="related">Related Services</h3>
       <ul>
         <li>
-          <a href="http://xip.io/">xip.io</a>: the inspiration for sslip.io
+          <a href="http://xip.io/">xip.io</a>: the inspiration for sslip.io. Sadly, this appears to be no longer
+          maintained after <a href="https://twitter.com/sstephenson/status/1388146129284603906">Sam Stephenson left
+          Basecamp</a>.
         </li>
         <li>
           <a href="http://nip.io">nip.io</a>: similar to xip.io, but the PowerDNS backend is written in elegant Python