Merge pull request #124 from up9inc/mizu_telemetry

Adding telemetry reports

.github/workflows/publish-docker.yml

@@ -1,39 +0,0 @@
-name: publish-docker
-on:
-  push:
-    branches:
-      - 'develop'
-      - 'main'
-jobs:
-  docker:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v2
-      - name: Get base image name
-        shell: bash
-        run: echo "##[set-output name=image;]$(echo gcr.io/up9-docker-hub/mizu/${GITHUB_REF#refs/heads/})"
-        id: base_image_step
-      - name: Docker meta
-        id: meta
-        uses: crazy-max/ghaction-docker-meta@v2
-        with:
-          images: ${{ steps.base_image_step.outputs.image }}
-          tags: |
-            type=sha
-            type=raw,${{ github.sha }}
-            type=raw,latest
-      - name: Login to DockerHub
-        uses: docker/login-action@v1
-        with:
-          registry: gcr.io
-          username: _json_key
-          password: ${{ secrets.GCR_JSON_KEY }}
-      - name: Build and push
-        uses: docker/build-push-action@v2
-        with:
-          context: .
-          push: true
-          tags: ${{ steps.meta.outputs.tags }}
-          labels: ${{ steps.meta.outputs.labels }}
-

.github/workflows/publish.yml

@@ -1,9 +1,9 @@
-name: public-cli
+name: publish
 on:
   push:
     branches:
-      - develop
-      - main
+      - 'develop'
+      - 'main'
 jobs:
   docker:
     runs-on: ubuntu-latest
@@ -27,12 +27,43 @@ jobs:
         with:
           releaseType: ${{ steps.condval.outputs.value }}
           github_token: ${{ secrets.GITHUB_TOKEN }}
-      - name: Get base image name
+      - name: Get version parameters
         shell: bash
         run: |
           echo "##[set-output name=build_timestamp;]$(echo $(date +%s))"
           echo "##[set-output name=branch;]$(echo ${GITHUB_REF#refs/heads/})"
         id: version_parameters
+      - name: Get base image name
+        shell: bash
+        run: echo "##[set-output name=image;]$(echo gcr.io/up9-docker-hub/mizu/${GITHUB_REF#refs/heads/})"
+        id: base_image_step
+      - name: Docker meta
+        id: meta
+        uses: crazy-max/ghaction-docker-meta@v2
+        with:
+          images: ${{ steps.base_image_step.outputs.image }}
+          tags: |
+            type=sha
+            type=raw,${{ github.sha }}
+            type=raw,${{ steps.versioning.outputs.version }}
+      - name: Login to DockerHub
+        uses: docker/login-action@v1
+        with:
+          registry: gcr.io
+          username: _json_key
+          password: ${{ secrets.GCR_JSON_KEY }}
+      - name: Build and push
+        uses: docker/build-push-action@v2
+        with:
+          context: .
+          push: true
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
+          build-args: |
+            SEM_VER=${{ steps.versioning.outputs.version }}
+            BUILD_TIMESTAMP=${{ steps.version_parameters.outputs.build_timestamp }}
+            GIT_BRANCH=${{ steps.version_parameters.outputs.branch }}
+            COMMIT_HASH=${{ github.sha }}
       - name: Build and Push CLI
         run: make push-cli SEM_VER='${{ steps.versioning.outputs.version }}' BUILD_TIMESTAMP='${{ steps.version_parameters.outputs.build_timestamp }}'
       - name: publish

Dockerfile

@@ -23,11 +23,20 @@ RUN go mod download
 # cheap trick to make the build faster (As long as go.mod wasn't changes)
 RUN go list -f '{{.Path}}@{{.Version}}' -m all | sed 1d | grep -e 'go-cache' -e 'sqlite' | xargs go get
 
+ARG COMMIT_HASH
+ARG GIT_BRANCH
+ARG BUILD_TIMESTAMP
+ARG SEM_VER
+
 # Copy and build api code
 COPY shared ../shared
 COPY tap ../tap
 COPY api .
-RUN go build -ldflags="-s -w" -o mizuagent .
+RUN go build -ldflags="-s -w \
+     -X 'mizuserver/pkg/version.GitCommitHash=${COMMIT_HASH}' \
+     -X 'mizuserver/pkg/version.Branch=${GIT_BRANCH}' \
+     -X 'mizuserver/pkg/version.BuildTimestamp=${BUILD_TIMESTAMP}' \
+     -X 'mizuserver/pkg/version.SemVer=${SEM_VER}'" -o mizuagent .
 
 
 FROM alpine:3.13.5

README.md

@@ -26,6 +26,49 @@ SHA256 checksums are available on the [Releases](https://github.com/up9inc/mizu/
 ### Development (unstable) build
 Pick one from the [Releases](https://github.com/up9inc/mizu/releases) page.
 
+## Prerequisites
+1. Set `KUBECONFIG` environment variable to your kubernetes configuration. If this is not set, mizu assumes that configuration is at `${HOME}/.kube/config`
+2. mizu needs following permissions on your kubernetes cluster to run
+```
+- apiGroups:
+  - ""
+  - apps
+  resources:
+  - pods
+  - services
+  verbs:
+  - list
+  - get
+  - create
+  - delete
+- apiGroups:
+  - ""
+  - apps
+  resources:
+  - daemonsets
+  verbs:
+  - list
+  - get
+  - create
+  - patch
+  - delete
+```
+3. Optionally, for resolving traffic ip to kubernetes service name, mizu needs below permissions
+```
+- apiGroups:
+  - ""
+  - apps
+  - "rbac.authorization.k8s.io"
+  resources:
+  - clusterroles
+  - clusterrolebindings
+  - serviceaccounts
+  verbs:
+  - get
+  - create
+  - delete
+```
+
 ## How to run
 
 1. Find pod you'd like to tap to in your Kubernetes cluster

api/main.go

@@ -101,6 +101,7 @@ func hostApi(socketHarOutputChannel chan<- *tap.OutputChannelItem) {
 	}
 	routes.WebSocketRoutes(app, &eventHandlers)
 	routes.EntriesRoutes(app)
+	routes.MetadataRoutes(app)
 	routes.NotFoundRoute(app)
 
 	utils.StartServer(app)

api/pkg/api/main.go

@@ -161,7 +161,7 @@ func saveHarToDb(entry *har.Entry, connectionInfo *tap.ConnectionInfo) {
 		IsOutgoing:          connectionInfo.IsOutgoing,
 	}
 	mizuEntry.EstimatedSizeBytes = getEstimatedEntrySizeBytes(mizuEntry)
-	database.GetEntriesTable().Create(&mizuEntry)
+	database.CreateEntry(&mizuEntry)
 
 	baseEntry := models.BaseEntryDetails{}
 	if err := models.GetEntry(&mizuEntry, &baseEntry); err != nil {

api/pkg/controllers/metadata_controller.go

@@ -0,0 +1,12 @@
+package controllers
+
+import (
+	"github.com/gofiber/fiber/v2"
+	"github.com/up9inc/mizu/shared"
+	"mizuserver/pkg/version"
+)
+
+func GetVersion(c *fiber.Ctx) error {
+	resp := shared.VersionResponse{SemVer: version.SemVer}
+	return c.Status(fiber.StatusOK).JSON(resp)
+}

api/pkg/database/main.go

@@ -12,11 +12,6 @@ import (
 
 const (
 	DBPath = "./entries.db"
-)
-
- var DB *gorm.DB
-
-const (
 	OrderDesc = "desc"
 	OrderAsc  = "asc"
 	LT        = "lt"
@@ -24,6 +19,8 @@ const (
 )
 
 var (
+	DB *gorm.DB
+	IsDBLocked = false
 	OperatorToSymbolMapping = map[string]string{
 		LT: "<",
 		GT: ">",
@@ -43,6 +40,13 @@ func GetEntriesTable() *gorm.DB {
 	return DB.Table("mizu_entries")
 }
 
+func CreateEntry(entry *models.MizuEntry) {
+	if IsDBLocked {
+		return
+	}
+	GetEntriesTable().Create(entry)
+}
+
 func initDataBase(databasePath string) *gorm.DB {
 	temp, _ := gorm.Open(sqlite.Open(databasePath), &gorm.Config{
 		Logger: &utils.TruncatingLogger{LogLevel: logger.Warn, SlowThreshold: 500 * time.Millisecond},

api/pkg/database/size_enforcer.go

@@ -81,6 +81,10 @@ func checkFileSize(maxSizeBytes int64) {
 }
 
 func pruneOldEntries(currentFileSize int64) {
+	// sqlite locks the database while delete or VACUUM are running and sqlite is terrible at handling its own db lock while a lot of inserts are attempted, we prevent a significant bottleneck by handling the db lock ourselves here
+	IsDBLocked = true
+	defer func() {IsDBLocked = false}()
+
 	amountOfBytesToTrim := currentFileSize / (100 / percentageOfMaxSizeBytesToPrune)
 
 	rows, err := GetEntriesTable().Limit(10000).Order("id").Rows()

api/pkg/resolver/loader.go

@@ -10,6 +10,7 @@ import (
 	"k8s.io/client-go/tools/clientcmd"
 	"k8s.io/client-go/util/homedir"
 	"path/filepath"
+	"os"
 )
 
 func NewFromInCluster(errOut chan error) (*Resolver, error) {
@@ -26,8 +27,13 @@ func NewFromInCluster(errOut chan error) (*Resolver, error) {
 
 func NewFromOutOfCluster(kubeConfigPath string, errOut chan error) (*Resolver, error) {
 	if kubeConfigPath == "" {
-		home := homedir.HomeDir()
-		kubeConfigPath = filepath.Join(home, ".kube", "config")
+		env := os.Getenv("KUBECONFIG") 
+		if env != "" {
+			kubeConfigPath = env
+		} else {
+			home := homedir.HomeDir()
+			kubeConfigPath = filepath.Join(home, ".kube", "config")
+		}
 	}
 
 	configPathList := filepath.SplitList(kubeConfigPath)

api/pkg/routes/entries_routes.go

@@ -5,7 +5,7 @@ import (
 	"mizuserver/pkg/controllers"
 )
 
-// EntriesRoutes func for describe group of public routes.
+// EntriesRoutes defines the group of har entries routes.
 func EntriesRoutes(fiberApp *fiber.App) {
 	routeGroup := fiberApp.Group("/api")
 

api/pkg/routes/metadata_routes.go

@@ -0,0 +1,13 @@
+package routes
+
+import (
+	"github.com/gofiber/fiber/v2"
+	"mizuserver/pkg/controllers"
+)
+
+// MetadataRoutes defines the group of metadata routes.
+func MetadataRoutes(fiberApp *fiber.App) {
+	routeGroup := fiberApp.Group("/metadata")
+
+	routeGroup.Get("/version", controllers.GetVersion)
+}

api/pkg/routes/not_found_route.go

@@ -2,7 +2,7 @@ package routes
 
 import "github.com/gofiber/fiber/v2"
 
-// NotFoundRoute func for describe 404 Error route.
+// NotFoundRoute defines the 404 Error route.
 func NotFoundRoute(fiberApp *fiber.App) {
 	fiberApp.Use(
 		func(c *fiber.Ctx) error {

api/pkg/version/consts.go

@@ -0,0 +1,8 @@
+package version
+
+var (
+	SemVer         = "0.0.1"
+	Branch         = "develop"
+	GitCommitHash  = "" // this var is overridden using ldflags in makefile when building
+	BuildTimestamp = "" // this var is overridden using ldflags in makefile when building
+)

build-push-featurebranch.sh

@@ -14,7 +14,7 @@ then
 fi
 
 echo "building $DOCKER_TAGGED_BUILD"
-docker build -t "$DOCKER_TAGGED_BUILD" .
+docker build -t "$DOCKER_TAGGED_BUILD" --build-arg SEM_VER=${SEM_VER} --build-arg BUILD_TIMESTAMP=${BUILD_TIMESTAMP} --build-arg GIT_BRANCH=${GIT_BRANCH} --build-arg COMMIT_HASH=${COMMIT_HASH} .
 
 echo pushing to "$REPOSITORY"
 docker push "$DOCKER_TAGGED_BUILD"

cli/cmd/fetch.go

@@ -2,6 +2,7 @@ package cmd
 
 import (
 	"github.com/spf13/cobra"
+	"github.com/up9inc/mizu/cli/mizu"
 )
 
 type MizuFetchOptions struct {
@@ -17,6 +18,12 @@ var fetchCmd = &cobra.Command{
 	Use:   "fetch",
 	Short: "Download recorded traffic to files",
 	RunE: func(cmd *cobra.Command, args []string) error {
+		go mizu.ReportRun("tap", mizuTapOptions)
+		if isCompatible, err := mizu.CheckVersionCompatibility(mizuFetchOptions.MizuPort); err != nil {
+			return err
+		} else if !isCompatible {
+			return nil
+		}
 		RunMizuFetch(&mizuFetchOptions)
 		return nil
 	},

cli/cmd/tap.go

@@ -31,8 +31,8 @@ var mizuTapOptions = &MizuTapOptions{}
 var direction string
 var humanMaxEntriesDBSize string
 var regex *regexp.Regexp
-const maxEntriesDBSizeFlagName = "max-entries-db-size"
 
+const maxEntriesDBSizeFlagName = "max-entries-db-size"
 
 const analysisMessageToConfirm = `NOTE: running mizu with --analysis flag will upload recorded traffic
 to UP9 cloud for further analysis and enriched presentation options.
@@ -44,6 +44,7 @@ var tapCmd = &cobra.Command{
 	Long: `Record the ingoing traffic of a kubernetes pod.
 Supported protocols are HTTP and gRPC.`,
 	RunE: func(cmd *cobra.Command, args []string) error {
+		go mizu.ReportRun("tap", mizuTapOptions)
 		RunMizuTap(regex, mizuTapOptions)
 		return nil
 	},
@@ -64,10 +65,8 @@ Supported protocols are HTTP and gRPC.`,
 		mizuTapOptions.MaxEntriesDBSizeBytes, parseHumanDataSizeErr = units.HumanReadableToBytes(humanMaxEntriesDBSize)
 		if parseHumanDataSizeErr != nil {
 			return errors.New(fmt.Sprintf("Could not parse --max-entries-db-size value %s", humanMaxEntriesDBSize))
-		} else if cmd.Flags().Changed(maxEntriesDBSizeFlagName) {
-			// We're parsing human readable file sizes here so its best to be unambiguous
-			fmt.Printf("Setting max entries db size to %s\n", units.BytesToHumanReadable(mizuTapOptions.MaxEntriesDBSizeBytes))
 		}
+		fmt.Printf("Mizu will store up to %s of traffic, old traffic will be cleared once the limit is reached.\n", units.BytesToHumanReadable(mizuTapOptions.MaxEntriesDBSizeBytes))
 
 		directionLowerCase := strings.ToLower(direction)
 		if directionLowerCase == "any" {
@@ -99,7 +98,7 @@ func init() {
 	tapCmd.Flags().Uint16VarP(&mizuTapOptions.SleepIntervalSec, "upload-interval", "", 10, "Interval in seconds for uploading data to UP9")
 	tapCmd.Flags().BoolVarP(&mizuTapOptions.AllNamespaces, "all-namespaces", "A", false, "Tap all namespaces")
 	tapCmd.Flags().StringVarP(&mizuTapOptions.KubeConfigPath, "kube-config", "k", "", "Path to kube-config file")
-	tapCmd.Flags().StringVarP(&mizuTapOptions.MizuImage, "mizu-image", "", fmt.Sprintf("gcr.io/up9-docker-hub/mizu/%s:latest", mizu.Branch), "Custom image for mizu collector")
+	tapCmd.Flags().StringVarP(&mizuTapOptions.MizuImage, "mizu-image", "", fmt.Sprintf("gcr.io/up9-docker-hub/mizu/%s:%s", mizu.Branch, mizu.SemVer), "Custom image for mizu collector")
 	tapCmd.Flags().StringArrayVarP(&mizuTapOptions.PlainTextFilterRegexes, "regex-masking", "r", nil, "List of regex expressions that are used to filter matching values from text/plain http bodies")
 	tapCmd.Flags().StringVarP(&direction, "direction", "", "in", "Record traffic that goes in this direction (relative to the tapped pod): in/any")
 	tapCmd.Flags().BoolVar(&mizuTapOptions.HideHealthChecks, "hide-healthchecks", false, "hides requests with kube-probe or prometheus user-agent headers")

cli/cmd/version.go

@@ -10,20 +10,20 @@ import (
 )
 
 type MizuVersionOptions struct {
-	DebugInfo                bool
+	DebugInfo bool
 }
 
-
 var mizuVersionOptions = &MizuVersionOptions{}
 
 var versionCmd = &cobra.Command{
 	Use:   "version",
 	Short: "Print version info",
 	RunE: func(cmd *cobra.Command, args []string) error {
+		go mizu.ReportRun("version", mizuVersionOptions)
 		if mizuVersionOptions.DebugInfo {
 			timeStampInt, _ := strconv.ParseInt(mizu.BuildTimestamp, 10, 0)
 			fmt.Printf("Version: %s \nBranch: %s (%s) \n", mizu.SemVer, mizu.Branch, mizu.GitCommitHash)
-			fmt.Printf("Build Time: %s (%s)\n", mizu.BuildTimestamp,  time.Unix(timeStampInt, 0))
+			fmt.Printf("Build Time: %s (%s)\n", mizu.BuildTimestamp, time.Unix(timeStampInt, 0))
 
 		} else {
 			fmt.Printf("Version: %s (%s)\n", mizu.SemVer, mizu.Branch)

cli/cmd/view.go

@@ -2,10 +2,11 @@ package cmd
 
 import (
 	"github.com/spf13/cobra"
+	"github.com/up9inc/mizu/cli/mizu"
 )
 
 type MizuViewOptions struct {
-	GuiPort                uint16
+	GuiPort uint16
 }
 
 var mizuViewOptions = &MizuViewOptions{}
@@ -14,6 +15,12 @@ var viewCmd = &cobra.Command{
 	Use:   "view",
 	Short: "Open GUI in browser",
 	RunE: func(cmd *cobra.Command, args []string) error {
+		go mizu.ReportRun("view", mizuFetchOptions)
+		if isCompatible, err := mizu.CheckVersionCompatibility(mizuFetchOptions.MizuPort); err != nil {
+			return err
+		} else if !isCompatible {
+			return nil
+		}
 		runMizuView(mizuViewOptions)
 		return nil
 	},

cli/mizu/telemetry.go

@@ -0,0 +1,30 @@
+package mizu
+
+import (
+	"bytes"
+	"encoding/json"
+	"fmt"
+	"github.com/romana/rlog"
+	"net/http"
+)
+
+const telemetryUrl = "https://us-east4-up9-prod.cloudfunctions.net/mizu-telemetry"
+
+func ReportRun(cmd string, args interface{}) {
+	if Branch != "main" {
+		rlog.Debugf("reporting only on main branch")
+		return
+	}
+
+	argsBytes, _ := json.Marshal(args)
+	argsMap := map[string]string{"telemetry_type": "mizu_execution", "cmd": cmd, "args": string(argsBytes), "component": "mizu_cli"}
+	argsMap["message"] = fmt.Sprintf("mizu %v - %v", argsMap["cmd"], string(argsBytes))
+
+	jsonValue, _ := json.Marshal(argsMap)
+
+	if resp, err := http.Post(telemetryUrl, "application/json", bytes.NewBuffer(jsonValue)); err != nil {
+		rlog.Debugf("error sending telemtry err: %v, response %v", err, resp)
+	} else {
+		rlog.Debugf("Successfully reported telemetry")
+	}
+}

cli/mizu/versionCheck.go

@@ -0,0 +1,46 @@
+package mizu
+
+import (
+	"encoding/json"
+	"fmt"
+	"github.com/up9inc/mizu/shared"
+	"github.com/up9inc/mizu/shared/semver"
+	"net/http"
+	"net/url"
+)
+
+
+func getApiVersion(port uint16) (string, error) {
+	versionUrl, _ := url.Parse(fmt.Sprintf("http://localhost:%d/mizu/metadata/version", port))
+	req := &http.Request{
+		Method: http.MethodGet,
+		URL:    versionUrl,
+	}
+	statusResp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		return "", err
+	}
+	defer statusResp.Body.Close()
+
+	versionResponse := &shared.VersionResponse{}
+	if err := json.NewDecoder(statusResp.Body).Decode(&versionResponse); err != nil {
+		return "", err
+	}
+
+	return versionResponse.SemVer, nil
+}
+
+func CheckVersionCompatibility(port uint16) (bool, error) {
+	apiSemVer, err := getApiVersion(port)
+	if err != nil {
+		return false, err
+	}
+
+	if semver.SemVersion(apiSemVer).Major() == semver.SemVersion(SemVer).Major() &&
+		semver.SemVersion(apiSemVer).Minor() == semver.SemVersion(SemVer).Minor() {
+		return true, nil
+	}
+
+	fmt.Printf(Red, fmt.Sprintf("cli version (%s) is not compatible with api version (%s)\n", SemVer, apiSemVer))
+	return false, nil
+}

shared/models.go

@@ -61,3 +61,8 @@ type TrafficFilteringOptions struct {
 	PlainTextMaskingRegexes []*SerializableRegexp
 	HideHealthChecks        bool
 }
+
+type VersionResponse struct {
+	SemVer string `json:"semver"`
+}
+

shared/semver/semver.go

@@ -0,0 +1,28 @@
+package semver
+
+import (
+	"regexp"
+)
+
+type SemVersion string
+
+func (v SemVersion) Breakdown() (string, string, string) {
+	re := regexp.MustCompile(`\d+`)
+	breakdown := re.FindAllString(string(v), 3)
+	return breakdown[0], breakdown[1], breakdown[2]
+}
+
+func (v SemVersion) Major() string {
+	major, _, _ := v.Breakdown()
+	return major
+}
+
+func (v SemVersion) Minor() string {
+	_, minor, _ := v.Breakdown()
+	return minor
+}
+
+func (v SemVersion) Patch() string {
+	_, _, patch := v.Breakdown()
+	return patch
+}

tap/har_writer.go

@@ -1,9 +1,11 @@
 package tap
 
 import (
+	"bytes"
 	"encoding/json"
 	"errors"
 	"fmt"
+	"io/ioutil"
 	"log"
 	"net/http"
 	"os"
@@ -46,12 +48,27 @@ type HarFile struct {
 }
 
 func NewEntry(request *http.Request, requestTime time.Time, response *http.Response, responseTime time.Time) (*har.Entry, error) {
-	harRequest, err := har.NewRequest(request, true)
+	harRequest, err := har.NewRequest(request, false)
 	if err != nil {
 		SilentError("convert-request-to-har", "Failed converting request to HAR %s (%v,%+v)", err, err, err)
 		return nil, errors.New("Failed converting request to HAR")
 	}
 
+	// For requests with multipart/form-data or application/x-www-form-urlencoded Content-Type,
+	// martian/har will parse the request body and place the parameters in harRequest.PostData.Params
+	// instead of harRequest.PostData.Text (as the HAR spec requires it).
+	// Mizu currently only looks at PostData.Text. Therefore, instead of letting martian/har set the content of
+	// PostData, always copy the request body to PostData.Text.
+	if (request.ContentLength > 0) {
+		reqBody, err := ioutil.ReadAll(request.Body)
+		if err != nil {
+			SilentError("read-request-body", "Failed converting request to HAR %s (%v,%+v)", err, err, err)
+			return nil, errors.New("Failed reading request body")
+		}
+		request.Body = ioutil.NopCloser(bytes.NewReader(reqBody))
+		harRequest.PostData.Text = string(reqBody)
+	}
+
 	harResponse, err := har.NewResponse(response, true)
 	if err != nil {
 		SilentError("convert-response-to-har", "Failed converting response to HAR %s (%v,%+v)", err, err, err)