add ability to fail probes and fatal on broker errors

Signed-off-by: Vasiliy Tolstov <v.tolstov@unistack.org>

.gitea/workflows/job_lint.yml

@@ -0,0 +1,29 @@
+name: lint
+
+on:
+  pull_request:
+    types: [opened, reopened, synchronize]
+    branches:
+    - master
+    - v3
+    - v4
+
+jobs:
+  lint:
+    runs-on: ubuntu-latest
+    steps:
+    - name: checkout code
+      uses: actions/checkout@v4
+      with:
+        filter: 'blob:none'
+    - name: setup go
+      uses: actions/setup-go@v5
+      with:
+        cache-dependency-path: "**/*.sum"
+        go-version: 'stable' 
+    - name: setup deps
+      run: go get -v ./...
+    - name: run lint
+      uses: https://github.com/golangci/golangci-lint-action@v6
+      with:
+        version: 'latest'

.gitea/workflows/job_test.yml

@@ -0,0 +1,34 @@
+name: test
+
+on:
+  pull_request:
+    types: [opened, reopened, synchronize]
+    branches:
+    - master
+    - v3
+    - v4
+  push:
+    branches:
+    - master
+    - v3
+    - v4
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    steps:
+    - name: checkout code
+      uses: actions/checkout@v4
+      with:
+        filter: 'blob:none'
+    - name: setup go
+      uses: actions/setup-go@v5
+      with:
+        cache-dependency-path: "**/*.sum"
+        go-version: 'stable'
+    - name: setup deps
+      run: go get -v ./...
+    - name: run test
+      env:
+        INTEGRATION_TESTS: yes
+      run: go test -mod readonly -v ./...

.gitea/workflows/job_tests.yml

@@ -0,0 +1,53 @@
+name: test
+
+on:
+  pull_request:
+    types: [opened, reopened, synchronize]
+    branches:
+    - master
+    - v3
+    - v4
+  push:
+    branches:
+    - master
+    - v3
+    - v4
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    steps:
+    - name: checkout code
+      uses: actions/checkout@v4
+      with:
+        filter: 'blob:none'
+    - name: checkout tests
+      uses: actions/checkout@v4
+      with:
+        ref: master
+        filter: 'blob:none'
+        repository: unistack-org/micro-tests
+        path: micro-tests
+    - name: setup go
+      uses: actions/setup-go@v5
+      with:
+        cache-dependency-path: "**/*.sum"
+        go-version: 'stable'
+    - name: setup go work
+      env:
+        GOWORK: /workspace/${{ github.repository_owner }}/go.work
+      run: |
+        go work init
+        go work use .
+        go work use micro-tests        
+    - name: setup deps
+      env:
+        GOWORK: /workspace/${{ github.repository_owner }}/go.work
+      run: go get -v ./...
+    - name: run tests
+      env:
+        INTEGRATION_TESTS: yes
+        GOWORK: /workspace/${{ github.repository_owner }}/go.work
+      run: |
+        cd micro-tests
+        go test -mod readonly -v ./... || true        

.github/dependabot.yml

@@ -1,19 +0,0 @@
-# To get started with Dependabot version updates, you'll need to specify which
-# package ecosystems to update and where the package manifests are located.
-# Please see the documentation for all configuration options:
-# https://help.github.com/github/administering-a-repository/configuration-options-for-dependency-updates
-
-version: 2
-updates:
-
-  # Maintain dependencies for GitHub Actions
-  - package-ecosystem: "github-actions"
-    directory: "/"
-    schedule:
-      interval: "daily"
-
-  # Maintain dependencies for Golang
-  - package-ecosystem: "gomod"
-    directory: "/"
-    schedule:
-      interval: "daily"

.github/workflows/autoapprove.yml

@@ -1,20 +0,0 @@
-name: "autoapprove"
-
-on:
-  pull_request_target:
-    types: [assigned, opened, synchronize, reopened]
-
-permissions:
-  pull-requests: write
-  contents: write
-
-jobs:
-  autoapprove:
-    runs-on: ubuntu-latest
-    steps:
-      - name: approve
-        uses: hmarr/auto-approve-action@v3
-        if: github.actor == 'vtolstov' || github.actor == 'dependabot[bot]'
-        id: approve
-        with:
-          github-token: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/automerge.yml

@@ -1,21 +0,0 @@
-name: "automerge"
-
-on:
-  pull_request_target:
-    types: [assigned, opened, synchronize, reopened]
-
-permissions:
-  pull-requests: write
-  contents: write
-
-jobs:
-  automerge:
-    runs-on: ubuntu-latest
-    if: github.actor == 'vtolstov'
-    steps:
-      - name: merge
-        id: merge
-        run: gh pr merge --auto --merge "$PR_URL"
-        env:
-          PR_URL: ${{github.event.pull_request.html_url}}
-          GITHUB_TOKEN: ${{secrets.TOKEN}}

.github/workflows/build.yml

@@ -1,47 +0,0 @@
-name: build
-on:
- push:
-    branches:
-    - master
-    - v3
-jobs:
-  test:
-    name: test
-    runs-on: ubuntu-latest
-    steps:
-    - name: setup
-      uses: actions/setup-go@v3
-      with:
-        go-version: 1.17
-    - name: checkout
-      uses: actions/checkout@v3
-    - name: cache
-      uses: actions/cache@v3
-      with:
-        path: ~/go/pkg/mod
-        key: ${{ runner.os }}-go-${{ hashFiles('**/go.sum') }}
-        restore-keys: ${{ runner.os }}-go-
-    - name: deps
-      run: go get -v -t -d ./...
-    - name: test
-      env:
-        INTEGRATION_TESTS: yes
-      run: go test -mod readonly -v ./...
-  lint:
-    name: lint
-    runs-on: ubuntu-latest
-    steps:
-      - name: checkout
-        uses: actions/checkout@v3
-      - name: lint
-        uses: golangci/golangci-lint-action@v3.4.0
-        continue-on-error: true
-        with:
-          # Required: the version of golangci-lint is required and must be specified without patch version: we always use the latest patch version.
-          version: v1.30
-          # Optional: working directory, useful for monorepos
-          # working-directory: somedir
-          # Optional: golangci-lint command line arguments.
-          # args: --issues-exit-code=0
-          # Optional: show only new issues if it's a pull request. The default value is `false`.
-          # only-new-issues: true

.github/workflows/codeql-analysis.yml

@@ -1,78 +0,0 @@
-# For most projects, this workflow file will not need changing; you simply need
-# to commit it to your repository.
-#
-# You may wish to alter this file to override the set of languages analyzed,
-# or to provide custom queries or build logic.
-#
-# ******** NOTE ********
-# We have attempted to detect the languages in your repository. Please check
-# the `language` matrix defined below to confirm you have the correct set of
-# supported CodeQL languages.
-#
-name: "codeql"
-
-on:
-  workflow_run:
-    workflows: ["prbuild"]
-    types:
-      - completed
-  push:
-    branches: [ master, v3 ]
-  pull_request:
-    # The branches below must be a subset of the branches above
-    branches: [ master, v3 ]
-  schedule:
-    - cron: '34 1 * * 0'
-
-jobs:
-  analyze:
-    name: analyze
-    runs-on: ubuntu-latest
-    permissions:
-      actions: read
-      contents: read
-      security-events: write
-
-    strategy:
-      fail-fast: false
-      matrix:
-        language: [ 'go' ]
-        # CodeQL supports [ 'cpp', 'csharp', 'go', 'java', 'javascript', 'python' ]
-        # Learn more:
-        # https://docs.github.com/en/free-pro-team@latest/github/finding-security-vulnerabilities-and-errors-in-your-code/configuring-code-scanning#changing-the-languages-that-are-analyzed
-
-    steps:
-    - name: checkout
-      uses: actions/checkout@v3
-    - name: setup
-      uses: actions/setup-go@v3
-      with:
-        go-version: 1.17
-    # Initializes the CodeQL tools for scanning.
-    - name: init
-      uses: github/codeql-action/init@v2
-      with:
-        languages: ${{ matrix.language }}
-        # If you wish to specify custom queries, you can do so here or in a config file.
-        # By default, queries listed here will override any specified in a config file.
-        # Prefix the list here with "+" to use these queries and those in the config file.
-        # queries: ./path/to/local/query, your-org/your-repo/queries@main
-
-    # Autobuild attempts to build any compiled languages  (C/C++, C#, or Java).
-    # If this step fails, then you should remove it and run the build manually (see below)
-    - name: autobuild
-      uses: github/codeql-action/autobuild@v2
-
-    # ℹ️ Command-line programs to run using the OS shell.
-    # 📚 https://git.io/JvXDl
-
-    # ✏️ If the Autobuild fails above, remove it and uncomment the following three lines
-    #    and modify them (or add more) to build your code if your project
-    #    uses a compiled language
-
-    #- run: |
-    #   make bootstrap
-    #   make release
-
-    - name: analyze
-      uses: github/codeql-action/analyze@v2

.github/workflows/dependabot-automerge.yml

@@ -1,27 +0,0 @@
-name: "dependabot-automerge"
-
-on:
-  pull_request_target:
-    types: [assigned, opened, synchronize, reopened]
-
-permissions:
-  pull-requests: write
-  contents: write
-
-jobs:
-  automerge:
-    runs-on: ubuntu-latest
-    if: github.actor == 'dependabot[bot]'
-    steps:
-      - name: metadata
-        id: metadata
-        uses: dependabot/fetch-metadata@v1.3.6
-        with:
-          github-token: "${{ secrets.TOKEN }}"
-      - name: merge
-        id: merge
-        if: ${{contains(steps.metadata.outputs.dependency-names, 'go.unistack.org')}}
-        run: gh pr merge --auto --merge "$PR_URL"
-        env:
-          PR_URL: ${{github.event.pull_request.html_url}}
-          GITHUB_TOKEN: ${{secrets.TOKEN}}

.github/workflows/pr.yml

@@ -1,47 +0,0 @@
-name: prbuild
-on:
-  pull_request:
-    branches:
-    - master
-    - v3
-jobs:
-  test:
-    name: test
-    runs-on: ubuntu-latest
-    steps:
-    - name: setup
-      uses: actions/setup-go@v3
-      with:
-        go-version: 1.17
-    - name: checkout
-      uses: actions/checkout@v3
-    - name: cache
-      uses: actions/cache@v3
-      with:
-        path: ~/go/pkg/mod
-        key: ${{ runner.os }}-go-${{ hashFiles('**/go.sum') }}
-        restore-keys: ${{ runner.os }}-go-
-    - name: deps
-      run: go get -v -t -d ./...
-    - name: test
-      env:
-        INTEGRATION_TESTS: yes
-      run: go test -mod readonly -v ./...
-  lint:
-    name: lint
-    runs-on: ubuntu-latest
-    steps:
-      - name: checkout
-        uses: actions/checkout@v3
-      - name: lint
-        uses: golangci/golangci-lint-action@v3.4.0
-        continue-on-error: true
-        with:
-          # Required: the version of golangci-lint is required and must be specified without patch version: we always use the latest patch version.
-          version: v1.30
-          # Optional: working directory, useful for monorepos
-          # working-directory: somedir
-          # Optional: golangci-lint command line arguments.
-          # args: --issues-exit-code=0
-          # Optional: show only new issues if it's a pull request. The default value is `false`.
-          # only-new-issues: true

.golangci.yml

@@ -0,0 +1,21 @@
+run:
+  concurrency: 8
+  deadline: 5m
+  issues-exit-code: 1
+  tests: true
+
+linters:
+  enable:
+    - staticcheck
+    - unused
+    - gosimple
+    - govet
+    - prealloc
+    - unconvert
+    - nakedret
+
+linters-settings:
+  govet:
+    check-all: true
+    enable:
+      - fieldalignment

README.md

@@ -1,2 +1,9 @@
 # broker-kgo
 
+TODO:
+
+1) экспортируем текущий оффсет для каждой партиции в группе
+2) экспортируем лаг для группы
+3) мониторим
+   1) если есть лаг больше нуля
+   2) если дельта оффсета за нужное нам время не 

broker.go

@@ -0,0 +1,69 @@
+package kgo
+
+import (
+	"context"
+	"net"
+	"sync/atomic"
+	"time"
+
+	"github.com/twmb/franz-go/pkg/kgo"
+	"go.unistack.org/micro/v3/logger"
+)
+
+type hookEvent struct {
+	log          logger.Logger
+	fatalOnError bool
+	connected    *atomic.Uint32
+}
+
+var (
+	_ kgo.HookBrokerConnect           = &hookEvent{}
+	_ kgo.HookBrokerDisconnect        = &hookEvent{}
+	_ kgo.HookBrokerRead              = &hookEvent{}
+	_ kgo.HookBrokerWrite             = &hookEvent{}
+	_ kgo.HookGroupManageError        = &hookEvent{}
+	_ kgo.HookProduceRecordUnbuffered = &hookEvent{}
+)
+
+func (m *hookEvent) OnGroupManageError(err error) {
+	if err != nil {
+		m.connected.Store(0)
+		if m.fatalOnError {
+			m.log.Fatal(context.TODO(), "kgo.OnGroupManageError", err)
+		}
+	}
+}
+
+func (m *hookEvent) OnBrokerConnect(_ kgo.BrokerMetadata, _ time.Duration, _ net.Conn, err error) {
+	if err != nil {
+		m.connected.Store(0)
+		if m.fatalOnError {
+			m.log.Fatal(context.TODO(), "kgo.OnBrokerConnect", err)
+		}
+	}
+}
+
+func (m *hookEvent) OnBrokerDisconnect(_ kgo.BrokerMetadata, _ net.Conn) {
+	m.connected.Store(0)
+}
+
+func (m *hookEvent) OnBrokerWrite(_ kgo.BrokerMetadata, _ int16, _ int, _ time.Duration, _ time.Duration, err error) {
+	if err != nil {
+		m.connected.Store(0)
+		if m.fatalOnError {
+			m.log.Fatal(context.TODO(), "kgo.OnBrokerWrite", err)
+		}
+	}
+}
+
+func (m *hookEvent) OnBrokerRead(_ kgo.BrokerMetadata, _ int16, _ int, _ time.Duration, _ time.Duration, err error) {
+	if err != nil {
+		m.connected.Store(0)
+	}
+}
+
+func (m *hookEvent) OnProduceRecordUnbuffered(_ *kgo.Record, err error) {
+	if err != nil {
+		m.connected.Store(0)
+	}
+}

carrier.go

@@ -1,6 +1,10 @@
 package kgo
 
 import (
+	"net/http"
+	"slices"
+	"strings"
+
 	"github.com/twmb/franz-go/pkg/kgo"
 	"go.unistack.org/micro/v3/metadata"
 )
@@ -53,24 +57,37 @@ func (c RecordCarrier) Keys() []string {
 	return out
 }
 
-func setHeaders(r *kgo.Record, md metadata.Metadata) {
+func setHeaders(r *kgo.Record, md metadata.Metadata, exclude ...string) {
 	seen := make(map[string]struct{})
+
 loop:
 	for k, v := range md {
+		k = http.CanonicalHeaderKey(k)
+
+		if _, ok := seen[k]; ok {
+			continue loop
+		}
+
+		if slices.ContainsFunc(exclude, func(s string) bool {
+			return strings.EqualFold(s, k)
+		}) {
+			continue loop
+		}
+
 		for i := 0; i < len(r.Headers); i++ {
-			if r.Headers[i].Key == k {
+			if strings.EqualFold(r.Headers[i].Key, k) {
 				// Key exist, update the value.
 				r.Headers[i].Value = []byte(v)
 				continue loop
-			} else if _, ok := seen[k]; ok {
-				continue loop
 			}
-			// Key does not exist, append new header.
-			r.Headers = append(r.Headers, kgo.RecordHeader{
-				Key:   k,
-				Value: []byte(v),
-			})
-			seen[k] = struct{}{}
 		}
+
+		// Key does not exist, append new header.
+		r.Headers = append(r.Headers, kgo.RecordHeader{
+			Key:   k,
+			Value: []byte(v),
+		})
+
+		seen[k] = struct{}{}
 	}
 }

event.go

@@ -8,11 +8,13 @@ import (
 )
 
 type event struct {
-	ctx   context.Context
-	topic string
-	err   error
-	sync.RWMutex
 	msg *broker.Message
+	err error
+	ctx context.Context
+
+	topic string
+
+	sync.RWMutex
 	ack bool
 }
 

go.mod

@@ -1,25 +1,27 @@
 module go.unistack.org/micro-broker-kgo/v3
 
-go 1.22.7
+go 1.23.0
-
-toolchain go1.23.3
 
 require (
 	github.com/google/uuid v1.6.0
-	github.com/twmb/franz-go v1.18.0
+	github.com/twmb/franz-go v1.18.1
-	github.com/twmb/franz-go/pkg/kadm v1.14.0
+	github.com/twmb/franz-go/pkg/kadm v1.15.0
 	github.com/twmb/franz-go/pkg/kmsg v1.9.0
-	go.opentelemetry.io/otel v1.32.0
+	go.opentelemetry.io/otel v1.34.0
-	go.unistack.org/micro/v3 v3.11.0
+	go.unistack.org/micro/v3 v3.11.41
 )
 
 require (
-	github.com/klauspost/compress v1.17.11 // indirect
+	github.com/ash3in/uuidv8 v1.2.0 // indirect
-	github.com/pierrec/lz4/v4 v4.1.21 // indirect
+	github.com/klauspost/compress v1.18.0 // indirect
+	github.com/kr/pretty v0.3.1 // indirect
+	github.com/matoous/go-nanoid v1.5.1 // indirect
+	github.com/pierrec/lz4/v4 v4.1.22 // indirect
 	go.unistack.org/micro-proto/v3 v3.4.1 // indirect
-	golang.org/x/crypto v0.29.0 // indirect
+	golang.org/x/crypto v0.35.0 // indirect
-	golang.org/x/sys v0.27.0 // indirect
+	golang.org/x/sys v0.30.0 // indirect
-	google.golang.org/genproto/googleapis/rpc v0.0.0-20241202173237-19429a94021a // indirect
+	google.golang.org/genproto/googleapis/rpc v0.0.0-20250224174004-546df14abb99 // indirect
-	google.golang.org/grpc v1.68.0 // indirect
+	google.golang.org/grpc v1.70.0 // indirect
-	google.golang.org/protobuf v1.35.2 // indirect
+	google.golang.org/protobuf v1.36.5 // indirect
+	gopkg.in/yaml.v3 v3.0.1 // indirect
 )

go.sum

@@ -1,28 +1,82 @@
+github.com/DATA-DOG/go-sqlmock v1.5.2 h1:OcvFkGmslmlZibjAjaHm3L//6LiuBgolP7OputlJIzU=
+github.com/DATA-DOG/go-sqlmock v1.5.2/go.mod h1:88MAG/4G7SMwSE3CeA0ZKzrT5CiOU3OJ+JlNzwDqpNU=
+github.com/ash3in/uuidv8 v1.2.0 h1:2oogGdtCPwaVtyvPPGin4TfZLtOGE5F+W++E880G6SI=
+github.com/ash3in/uuidv8 v1.2.0/go.mod h1:BnU0wJBxnzdEKmVg4xckBkD+VZuecTFTUP3M0dWgyY4=
+github.com/creack/pty v1.1.9/go.mod h1:oKZEueFk5CKHvIhNR5MUki03XCEU+Q6VDXinZuGJ33E=
+github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc h1:U9qPSI2PIWSS1VwoXQT9A3Wy9MM3WgvqSxFWenqJduM=
+github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
+github.com/golang/protobuf v1.5.4 h1:i7eJL8qZTpSEXOPTxNKhASYpMn+8e5Q6AdndVa1dWek=
+github.com/golang/protobuf v1.5.4/go.mod h1:lnTiLA8Wa4RWRcIUkrtSVa5nRhsEGBg48fD6rSs7xps=
+github.com/google/go-cmp v0.6.0 h1:ofyhxvXcZhMsU5ulbFiLKl/XBFqE1GSq7atu8tAmTRI=
+github.com/google/go-cmp v0.6.0/go.mod h1:17dUlkBOakJ0+DkrSSNjCkIjxS6bF9zb3elmeNGIjoY=
 github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
 github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
 github.com/klauspost/compress v1.17.11 h1:In6xLpyWOi1+C7tXUUWv2ot1QvBjxevKAaI6IXrJmUc=
 github.com/klauspost/compress v1.17.11/go.mod h1:pMDklpSncoRMuLFrf1W9Ss9KT+0rH90U12bZKk7uwG0=
-github.com/pierrec/lz4/v4 v4.1.21 h1:yOVMLb6qSIDP67pl/5F7RepeKYu/VmTyEXvuMI5d9mQ=
+github.com/klauspost/compress v1.18.0 h1:c/Cqfb0r+Yi+JtIEq73FWXVkRonBlf0CRNYc8Zttxdo=
-github.com/pierrec/lz4/v4 v4.1.21/go.mod h1:gZWDp/Ze/IJXGXf23ltt2EXimqmTUXEy0GFuRQyBid4=
+github.com/klauspost/compress v1.18.0/go.mod h1:2Pp+KzxcywXVXMr50+X0Q/Lsb43OQHYWRCY2AiWywWQ=
+github.com/kr/pretty v0.3.1 h1:flRD4NNwYAUpkphVc1HcthR4KEIFJ65n8Mw5qdRn3LE=
+github.com/kr/pretty v0.3.1/go.mod h1:hoEshYVHaxMs3cyo3Yncou5ZscifuDolrwPKZanG3xk=
+github.com/kr/text v0.2.0 h1:5Nx0Ya0ZqY2ygV366QzturHI13Jq95ApcVaJBhpS+AY=
+github.com/kr/text v0.2.0/go.mod h1:eLer722TekiGuMkidMxC/pM04lWEeraHUUmBw8l2grE=
+github.com/matoous/go-nanoid v1.5.1 h1:aCjdvTyO9LLnTIi0fgdXhOPPvOHjpXN6Ik9DaNjIct4=
+github.com/matoous/go-nanoid v1.5.1/go.mod h1:zyD2a71IubI24efhpvkJz+ZwfwagzgSO6UNiFsZKN7U=
+github.com/pierrec/lz4/v4 v4.1.22 h1:cKFw6uJDK+/gfw5BcDL0JL5aBsAFdsIT18eRtLj7VIU=
+github.com/pierrec/lz4/v4 v4.1.22/go.mod h1:gZWDp/Ze/IJXGXf23ltt2EXimqmTUXEy0GFuRQyBid4=
+github.com/pkg/diff v0.0.0-20210226163009-20ebb0f2a09e/go.mod h1:pJLUxLENpZxwdsKMEsNbx1VGcRFpLqf3715MtcvvzbA=
+github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2 h1:Jamvg5psRIccs7FGNTlIRMkT8wgtp5eCXdBlqhYGL6U=
+github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
+github.com/rogpeppe/go-internal v1.9.0/go.mod h1:WtVeX8xhTBvf0smdhujwtBcq4Qrzq/fJaraNFVN+nFs=
+github.com/rogpeppe/go-internal v1.13.1 h1:KvO1DLK/DRN07sQ1LQKScxyZJuNnedQ5/wKSR38lUII=
+github.com/rogpeppe/go-internal v1.13.1/go.mod h1:uMEvuHeurkdAXX61udpOXGD/AzZDWNMNyH2VO9fmH0o=
+github.com/stretchr/testify v1.10.0 h1:Xv5erBjTwe/5IxqUQTdXv5kgmIvbHo3QQyRwhJsOfJA=
+github.com/stretchr/testify v1.10.0/go.mod h1:r2ic/lqez/lEtzL7wO/rwa5dbSLXVDPFyf8C91i36aY=
 github.com/twmb/franz-go v1.18.0 h1:25FjMZfdozBywVX+5xrWC2W+W76i0xykKjTdEeD2ejw=
 github.com/twmb/franz-go v1.18.0/go.mod h1:zXCGy74M0p5FbXsLeASdyvfLFsBvTubVqctIaa5wQ+I=
+github.com/twmb/franz-go v1.18.1 h1:D75xxCDyvTqBSiImFx2lkPduE39jz1vaD7+FNc+vMkc=
+github.com/twmb/franz-go v1.18.1/go.mod h1:Uzo77TarcLTUZeLuGq+9lNpSkfZI+JErv7YJhlDjs9M=
 github.com/twmb/franz-go/pkg/kadm v1.14.0 h1:nAn1co1lXzJQocpzyIyOFOjUBf4WHWs5/fTprXy2IZs=
 github.com/twmb/franz-go/pkg/kadm v1.14.0/go.mod h1:XjOPz6ZaXXjrW2jVCfLuucP8H1w2TvD6y3PT2M+aAM4=
+github.com/twmb/franz-go/pkg/kadm v1.15.0 h1:Yo3NAPfcsx3Gg9/hdhq4vmwO77TqRRkvpUcGWzjworc=
+github.com/twmb/franz-go/pkg/kadm v1.15.0/go.mod h1:MUdcUtnf9ph4SFBLLA/XxE29rvLhWYLM9Ygb8dfSCvw=
 github.com/twmb/franz-go/pkg/kmsg v1.9.0 h1:JojYUph2TKAau6SBtErXpXGC7E3gg4vGZMv9xFU/B6M=
 github.com/twmb/franz-go/pkg/kmsg v1.9.0/go.mod h1:CMbfazviCyY6HM0SXuG5t9vOwYDHRCSrJJyBAe5paqg=
-go.opentelemetry.io/otel v1.32.0 h1:WnBN+Xjcteh0zdk01SVqV55d/m62NJLJdIyb4y/WO5U=
+go.opentelemetry.io/otel v1.33.0 h1:/FerN9bax5LoK51X/sI0SVYrjSE0/yUL7DpxW4K3FWw=
-go.opentelemetry.io/otel v1.32.0/go.mod h1:00DCVSB0RQcnzlwyTfqtxSm+DRr9hpYrHjNGiBHVQIg=
+go.opentelemetry.io/otel v1.33.0/go.mod h1:SUUkR6csvUQl+yjReHu5uM3EtVV7MBm5FHKRlNx4I8I=
+go.opentelemetry.io/otel v1.34.0 h1:zRLXxLCgL1WyKsPVrgbSdMN4c0FMkDAskSTQP+0hdUY=
+go.opentelemetry.io/otel v1.34.0/go.mod h1:OWFPOQ+h4G8xpyjgqo4SxJYdDQ/qmRH+wivy7zzx9oI=
 go.unistack.org/micro-proto/v3 v3.4.1 h1:UTjLSRz2YZuaHk9iSlVqqsA50JQNAEK2ZFboGqtEa9Q=
 go.unistack.org/micro-proto/v3 v3.4.1/go.mod h1:okx/cnOhzuCX0ggl/vToatbCupi0O44diiiLLsZ93Zo=
-go.unistack.org/micro/v3 v3.11.0 h1:usQ+8wQuOWpQd4+DGhFXSgZ+e+wOBjuT3W5GJZ02bSs=
+go.unistack.org/micro/v3 v3.11.37 h1:ZcpnXAYEMcAwmnVb5b7o8/PylGnILxXMHaUlRrPmRI0=
-go.unistack.org/micro/v3 v3.11.0/go.mod h1:YzMldzHN9Ei+zy5t/Psu7RUWDZwUfrNYiStSQtTz90g=
+go.unistack.org/micro/v3 v3.11.37/go.mod h1:POGU5hstnAT9LH70m8FalyQSNi2GfIew71K75JenIZk=
-golang.org/x/crypto v0.29.0 h1:L5SG1JTTXupVV3n6sUqMTeWbjAyfPwoda2DLX8J8FrQ=
+go.unistack.org/micro/v3 v3.11.41 h1:dP4sBLIZpMo+MWGe5bbESewK8wBzYm4Yik/67x4dEtQ=
-golang.org/x/crypto v0.29.0/go.mod h1:+F4F4N5hv6v38hfeYwTdx20oUvLLc+QfrE9Ax9HtgRg=
+go.unistack.org/micro/v3 v3.11.41/go.mod h1:POGU5hstnAT9LH70m8FalyQSNi2GfIew71K75JenIZk=
-golang.org/x/sys v0.27.0 h1:wBqf8DvsY9Y/2P8gAfPDEYNuS30J4lPHJxXSb/nJZ+s=
+golang.org/x/crypto v0.31.0 h1:ihbySMvVjLAeSH1IbfcRTkD/iNscyz8rGzjF/E5hV6U=
-golang.org/x/sys v0.27.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
+golang.org/x/crypto v0.31.0/go.mod h1:kDsLvtWBEx7MV9tJOj9bnXsPbxwJQ6csT/x4KIN4Ssk=
-google.golang.org/genproto/googleapis/rpc v0.0.0-20241202173237-19429a94021a h1:hgh8P4EuoxpsuKMXX/To36nOFD7vixReXgn8lPGnt+o=
+golang.org/x/crypto v0.35.0 h1:b15kiHdrGCHrP6LvwaQ3c03kgNhhiMgvlhxHQhmg2Xs=
-google.golang.org/genproto/googleapis/rpc v0.0.0-20241202173237-19429a94021a/go.mod h1:5uTbfoYQed2U9p3KIj2/Zzm02PYhndfdmML0qC3q3FU=
+golang.org/x/crypto v0.35.0/go.mod h1:dy7dXNW32cAb/6/PRuTNsix8T+vJAqvuIy5Bli/x0YQ=
-google.golang.org/grpc v1.68.0 h1:aHQeeJbo8zAkAa3pRzrVjZlbz6uSfeOXlJNQM0RAbz0=
+golang.org/x/net v0.33.0 h1:74SYHlV8BIgHIFC/LrYkOGIwL19eTYXQ5wc6TBuO36I=
-google.golang.org/grpc v1.68.0/go.mod h1:fmSPC5AsjSBCK54MyHRx48kpOti1/jRfOlwEWywNjWA=
+golang.org/x/net v0.33.0/go.mod h1:HXLR5J+9DxmrqMwG9qjGCxZ+zKXxBru04zlTvWlWuN4=
-google.golang.org/protobuf v1.35.2 h1:8Ar7bF+apOIoThw1EdZl0p1oWvMqTHmpA2fRTyZO8io=
+golang.org/x/sys v0.28.0 h1:Fksou7UEQUWlKvIdsqzJmUmCX3cZuD2+P3XyyzwMhlA=
-google.golang.org/protobuf v1.35.2/go.mod h1:9fA7Ob0pmnwhb644+1+CVWFRbNajQ6iRojtC/QF5bRE=
+golang.org/x/sys v0.28.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
+golang.org/x/sys v0.30.0 h1:QjkSwP/36a20jFYWkSue1YwXzLmsV5Gfq7Eiy72C1uc=
+golang.org/x/sys v0.30.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
+golang.org/x/text v0.21.0 h1:zyQAAkrwaneQ066sspRyJaG9VNi/YJ1NfzcGB3hZ/qo=
+golang.org/x/text v0.21.0/go.mod h1:4IBbMaMmOPCJ8SecivzSH54+73PCFmPWxNTLm+vZkEQ=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20241216192217-9240e9c98484 h1:Z7FRVJPSMaHQxD0uXU8WdgFh8PseLM8Q8NzhnpMrBhQ=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20241216192217-9240e9c98484/go.mod h1:lcTa1sDdWEIHMWlITnIczmw5w60CF9ffkb8Z+DVmmjA=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20250224174004-546df14abb99 h1:ZSlhAUqC4r8TPzqLXQ0m3upBNZeF+Y8jQ3c4CR3Ujms=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20250224174004-546df14abb99/go.mod h1:LuRYeWDFV6WOn90g357N17oMCaxpgCnbi/44qJvDn2I=
+google.golang.org/grpc v1.69.2 h1:U3S9QEtbXC0bYNvRtcoklF3xGtLViumSYxWykJS+7AU=
+google.golang.org/grpc v1.69.2/go.mod h1:vyjdE6jLBI76dgpDojsFGNaHlxdjXN9ghpnd2o7JGZ4=
+google.golang.org/grpc v1.70.0 h1:pWFv03aZoHzlRKHWicjsZytKAiYCtNS0dHbXnIdq7jQ=
+google.golang.org/grpc v1.70.0/go.mod h1:ofIJqVKDXx/JiXrwr2IG4/zwdH9txy3IlF40RmcJSQw=
+google.golang.org/protobuf v1.36.1 h1:yBPeRvTftaleIgM3PZ/WBIZ7XM/eEYAaEyCwvyjq/gk=
+google.golang.org/protobuf v1.36.1/go.mod h1:9fA7Ob0pmnwhb644+1+CVWFRbNajQ6iRojtC/QF5bRE=
+google.golang.org/protobuf v1.36.5 h1:tPhr+woSbjfYvY6/GPufUoYizxw1cF/yFoxJ2fmpwlM=
+google.golang.org/protobuf v1.36.5/go.mod h1:9fA7Ob0pmnwhb644+1+CVWFRbNajQ6iRojtC/QF5bRE=
+gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
+gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c h1:Hei/4ADfdWqJk1ZMxUNpqntNwaWcugrBjAiHlqqRiVk=
+gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c/go.mod h1:JHkPIbrfpd72SG/EVd6muEfDQjcINNoR0C8j2r3qZ4Q=
+gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
+gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=

kadmtest.go

@@ -1,65 +0,0 @@
-//go:build ignore
-
-package main
-
-import (
-	"context"
-	"fmt"
-	"os"
-	"time"
-
-	"github.com/twmb/franz-go/pkg/kadm"
-	"github.com/twmb/franz-go/pkg/kgo"
-	"github.com/twmb/franz-go/pkg/kversion"
-
-	//"github.com/twmb/franz-go/pkg/sasl/scram"
-	"github.com/twmb/franz-go/pkg/sasl/plain"
-)
-
-func die(msg string, args ...any) {
-	fmt.Fprintf(os.Stderr, msg, args...)
-	os.Exit(1)
-}
-
-func main() {
-	seeds := []string{"vm-kafka-ump01tn.mbrd.ru:9092", "vm-kafka-ump02tn.mbrd.ru:9092", "vm-kafka-ump03tn.mbrd.ru:9092"}
-
-	pass := "XXXXX"
-	user := "XXXXX"
-
-	var adminClient *kadm.Client
-	{
-		client, err := kgo.NewClient(
-			kgo.SeedBrokers(seeds...),
-			// kgo.SASL((scram.Auth{User: user, Pass: pass}).AsSha512Mechanism()),
-			kgo.SASL((plain.Auth{User: user, Pass: pass}).AsMechanism()),
-
-			// Do not try to send requests newer than 2.4.0 to avoid breaking changes in the request struct.
-			// Sometimes there are breaking changes for newer versions where more properties are required to set.
-			kgo.MaxVersions(kversion.V2_4_0()),
-		)
-		if err != nil {
-			panic(err)
-		}
-		defer client.Close()
-
-		adminClient = kadm.NewClient(client)
-	}
-
-	ctx, cancel := context.WithTimeout(context.Background(), 15*time.Second)
-	defer cancel()
-
-	dg, err := adminClient.DescribeGroups(ctx, "interestrate_loader")
-	if err != nil {
-		die("failed to describe group: %v", err)
-	}
-
-	for _, m := range dg["interestrate_loader"].Members {
-		mc, _ := m.Assigned.AsConsumer()
-		for _, mt := range mc.Topics {
-			for _, p := range mt.Partitions {
-				fmt.Printf("client:%s\tpartitions: %d\n", m.ClientID, p)
-			}
-		}
-	}
-}

kgo.go

@@ -5,8 +5,7 @@ import (
 	"context"
 	"errors"
 	"fmt"
-	"math/rand"
+	"math/rand/v2"
-	"net/http"
 	"strings"
 	"sync"
 	"sync/atomic"
@@ -29,7 +28,6 @@ var ErrLostMessage = errors.New("message not marked for offsets commit and will
 
 var DefaultRetryBackoffFn = func() func(int) time.Duration {
 	var rngMu sync.Mutex
-	rng := rand.New(rand.NewSource(time.Now().UnixNano()))
 	return func(fails int) time.Duration {
 		const (
 			min = 100 * time.Millisecond
@@ -45,7 +43,7 @@ var DefaultRetryBackoffFn = func() func(int) time.Duration {
 		backoff := min * time.Duration(1<<(fails-1))
 
 		rngMu.Lock()
-		jitter := 0.8 + 0.4*rng.Float64()
+		jitter := 0.8 + 0.4*rand.Float64()
 		rngMu.Unlock()
 
 		backoff = time.Duration(float64(backoff) * jitter)
@@ -58,13 +56,16 @@ var DefaultRetryBackoffFn = func() func(int) time.Duration {
 }()
 
 type Broker struct {
-	init      bool
 	c         *kgo.Client
-	kopts     []kgo.Opt
 	connected *atomic.Uint32
-	sync.RWMutex
+
+	kopts []kgo.Opt
+	subs  []*Subscriber
+
 	opts broker.Options
-	subs []*Subscriber
+
+	sync.RWMutex
+	init bool
 }
 
 func (r *Broker) Live() bool {
@@ -108,10 +109,18 @@ func (k *Broker) connect(ctx context.Context, opts ...kgo.Opt) (*kgo.Client, *ho
 		}
 	}
 
+	var fatalOnError bool
+	if k.opts.Context != nil {
+		if v, ok := k.opts.Context.Value(fatalOnErrorKey{}).(bool); ok && v {
+			fatalOnError = v
+		}
+	}
+
 	htracer := &hookTracer{group: group, clientID: clientID, tracer: k.opts.Tracer}
 	opts = append(opts,
 		kgo.WithHooks(&hookMeter{meter: k.opts.Meter}),
 		kgo.WithHooks(htracer),
+		kgo.WithHooks(&hookEvent{log: k.opts.Logger, fatalOnError: fatalOnError, connected: k.connected}),
 	)
 
 	select {
@@ -258,7 +267,7 @@ func (k *Broker) publish(ctx context.Context, msgs []*broker.Message, opts ...br
 		k.connected.Store(1)
 	}
 	k.Unlock()
-
+	fmt.Printf("EEE\n")
 	options := broker.NewPublishOptions(opts...)
 	records := make([]*kgo.Record, 0, len(msgs))
 	var errs []string
@@ -277,14 +286,14 @@ func (k *Broker) publish(ctx context.Context, msgs []*broker.Message, opts ...br
 
 	for _, msg := range msgs {
 		rec := &kgo.Record{Context: ctx, Key: key}
+
 		rec.Topic, _ = msg.Header.Get(metadata.HeaderTopic)
 		msg.Header.Del(metadata.HeaderTopic)
+
 		k.opts.Meter.Counter(semconv.PublishMessageInflight, "endpoint", rec.Topic, "topic", rec.Topic).Inc()
 		if options.BodyOnly || k.opts.Codec.String() == "noop" {
 			rec.Value = msg.Body
-			for k, v := range msg.Header {
+			setHeaders(rec, msg.Header)
-				rec.Headers = append(rec.Headers, kgo.RecordHeader{Key: http.CanonicalHeaderKey(k), Value: []byte(v)})
-			}
 		} else {
 			rec.Value, err = k.opts.Codec.Marshal(msg)
 			if err != nil {
@@ -313,7 +322,9 @@ func (k *Broker) publish(ctx context.Context, msgs []*broker.Message, opts ...br
 		return nil
 	}
 	ts := time.Now()
+	fmt.Printf("SSSSSSEEE\n")
 	results := k.c.ProduceSync(ctx, records...)
+	fmt.Printf("SSSSSS\n")
 	te := time.Since(ts)
 	for _, result := range results {
 		k.opts.Meter.Summary(semconv.PublishMessageLatencyMicroseconds, "endpoint", result.Record.Topic, "topic", result.Record.Topic).Update(te.Seconds())
@@ -350,7 +361,7 @@ func (k *Broker) TopicExists(ctx context.Context, topic string) error {
 	return nil
 }
 
-func (k *Broker) BatchSubscribe(ctx context.Context, topic string, handler broker.BatchHandler, opts ...broker.SubscribeOption) (broker.Subscriber, error) {
+func (k *Broker) BatchSubscribe(_ context.Context, _ string, _ broker.BatchHandler, _ ...broker.SubscribeOption) (broker.Subscriber, error) {
 	return nil, nil
 }
 
@@ -372,13 +383,27 @@ func (k *Broker) Subscribe(ctx context.Context, topic string, handler broker.Han
 		}
 	}
 
+	var fatalOnError bool
+	if k.opts.Context != nil {
+		if v, ok := k.opts.Context.Value(fatalOnErrorKey{}).(bool); ok && v {
+			fatalOnError = v
+		}
+	}
+
+	if options.Context != nil {
+		if v, ok := options.Context.Value(fatalOnErrorKey{}).(bool); ok && v {
+			fatalOnError = v
+		}
+	}
+
 	sub := &Subscriber{
-		topic:     topic,
+		topic:        topic,
-		opts:      options,
+		opts:         options,
-		handler:   handler,
+		handler:      handler,
-		kopts:     k.opts,
+		kopts:        k.opts,
-		consumers: make(map[tp]*consumer),
+		consumers:    make(map[tp]*consumer),
-		done:      make(chan struct{}),
+		done:         make(chan struct{}),
+		fatalOnError: fatalOnError,
 	}
 
 	kopts := append(k.kopts,
@@ -389,7 +414,9 @@ func (k *Broker) Subscribe(ctx context.Context, topic string, handler broker.Han
 		kgo.AutoCommitInterval(commitInterval),
 		kgo.OnPartitionsAssigned(sub.assigned),
 		kgo.OnPartitionsRevoked(sub.revoked),
+		kgo.StopProducerOnDataLossDetected(),
 		kgo.OnPartitionsLost(sub.lost),
+		kgo.AutoCommitCallback(sub.autocommit),
 		kgo.AutoCommitMarks(),
 	)
 
@@ -450,7 +477,7 @@ func NewBroker(opts ...broker.Option) *Broker {
 		kgo.BlockRebalanceOnPoll(),
 		kgo.Balancers(kgo.CooperativeStickyBalancer()),
 		kgo.FetchIsolationLevel(kgo.ReadUncommitted()),
-		kgo.UnknownTopicRetries(0),
+		kgo.UnknownTopicRetries(1),
 	}
 
 	if options.Context != nil {

kgo_test.go

@@ -13,14 +13,15 @@ import (
 	kgo "go.unistack.org/micro-broker-kgo/v3"
 	"go.unistack.org/micro/v3/broker"
 	"go.unistack.org/micro/v3/logger"
+	"go.unistack.org/micro/v3/logger/slog"
 	"go.unistack.org/micro/v3/metadata"
 )
 
 var (
-	msgcnt   = int64(12000000)
+	msgcnt   = int64(1200)
 	group    = "38"
-	prefill  = false
+	prefill  = true
-	loglevel = logger.InfoLevel
+	loglevel = logger.DebugLevel
 )
 
 var bm = &broker.Message{
@@ -28,6 +29,94 @@ var bm = &broker.Message{
 	Body:   []byte(`"body"`),
 }
 
+func TestFail(t *testing.T) {
+	if tr := os.Getenv("INTEGRATION_TESTS"); len(tr) > 0 {
+		t.Skip()
+	}
+
+	logger.DefaultLogger = slog.NewLogger()
+	if err := logger.DefaultLogger.Init(logger.WithLevel(loglevel)); err != nil {
+		t.Fatal(err)
+	}
+	ctx := context.Background()
+
+	var addrs []string
+	if addr := os.Getenv("BROKER_ADDRS"); len(addr) == 0 {
+		addrs = []string{"127.0.0.1:9092"}
+	} else {
+		addrs = strings.Split(addr, ",")
+	}
+
+	b := kgo.NewBroker(
+		broker.Addrs(addrs...),
+		kgo.CommitInterval(5*time.Second),
+		kgo.Options(kg.ClientID("test"), kg.FetchMaxBytes(10*1024*1024),
+			kg.AllowAutoTopicCreation(),
+		),
+	)
+
+	t.Logf("broker init")
+	if err := b.Init(); err != nil {
+		t.Fatal(err)
+	}
+
+	t.Logf("broker connect")
+	if err := b.Connect(ctx); err != nil {
+		t.Fatal(err)
+	}
+
+	defer func() {
+		t.Logf("broker disconnect")
+		if err := b.Disconnect(ctx); err != nil {
+			t.Fatal(err)
+		}
+	}()
+
+	t.Logf("broker health %v", b.Health())
+	msgs := make([]*broker.Message, 0, msgcnt)
+	for i := int64(0); i < msgcnt; i++ {
+		msgs = append(msgs, bm)
+	}
+
+	for _, msg := range msgs {
+		t.Logf("broker publish")
+		if err := b.Publish(ctx, "test", msg); err != nil {
+			t.Fatal(err)
+		}
+	}
+	//	t.Skip()
+
+	idx := int64(0)
+	fn := func(msg broker.Event) error {
+		atomic.AddInt64(&idx, 1)
+		time.Sleep(500 * time.Millisecond)
+		t.Logf("ack")
+		return msg.Ack()
+	}
+
+	sub, err := b.Subscribe(ctx, "test", fn,
+		broker.SubscribeAutoAck(true),
+		broker.SubscribeGroup(group),
+		broker.SubscribeBodyOnly(true))
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer func() {
+		if err := sub.Unsubscribe(ctx); err != nil {
+			t.Fatal(err)
+		}
+	}()
+
+	for {
+		t.Logf("health check")
+		if !b.Health() {
+			t.Logf("health works")
+			break
+		}
+		time.Sleep(1 * time.Second)
+	}
+}
+
 func TestConnect(t *testing.T) {
 	var addrs []string
 	ctx := context.TODO()
@@ -67,6 +156,7 @@ func TestPubSub(t *testing.T) {
 		kgo.CommitInterval(5*time.Second),
 		kgo.Options(kg.ClientID("test"), kg.FetchMaxBytes(10*1024*1024)),
 	)
+
 	if err := b.Init(); err != nil {
 		t.Fatal(err)
 	}

meter.go

@@ -61,7 +61,7 @@ const (
 	labelTopic   = "topic"
 )
 
-func (m *hookMeter) OnGroupManageError(err error) {
+func (m *hookMeter) OnGroupManageError(_ error) {
 	m.meter.Counter(metricBrokerGroupErrors).Inc()
 }
 

options.go

@@ -72,6 +72,12 @@ func SubscribeOptions(opts ...kgo.Opt) broker.SubscribeOption {
 	}
 }
 
+type fatalOnErrorKey struct{}
+
+func FatalOnError(b bool) broker.Option {
+	return broker.SetOption(fatalOnErrorKey{}, b)
+}
+
 type clientIDKey struct{}
 
 func ClientID(id string) broker.Option {
@@ -98,6 +104,11 @@ func SubscribeMaxInFlight(n int) broker.SubscribeOption {
 	return broker.SetSubscribeOption(subscribeMaxInflightKey{}, n)
 }
 
+// SubscribeMaxInFlight max queued messages
+func SubscribeFatalOnError(b bool) broker.SubscribeOption {
+	return broker.SetSubscribeOption(fatalOnErrorKey{}, b)
+}
+
 type publishPromiseKey struct{}
 
 // PublishPromise set the kafka promise func for Produce

subscriber.go

@@ -5,10 +5,12 @@ import (
 	"fmt"
 	"strconv"
 	"sync"
+	"sync/atomic"
 	"time"
 
 	"github.com/twmb/franz-go/pkg/kadm"
 	"github.com/twmb/franz-go/pkg/kgo"
+	"github.com/twmb/franz-go/pkg/kmsg"
 	"go.unistack.org/micro/v3/broker"
 	"go.unistack.org/micro/v3/logger"
 	"go.unistack.org/micro/v3/metadata"
@@ -22,29 +24,34 @@ type tp struct {
 }
 
 type consumer struct {
-	c         *kgo.Client
 	topic     string
-	partition int32
+	c         *kgo.Client
 	htracer   *hookTracer
-	opts      broker.SubscribeOptions
-	kopts     broker.Options
-	handler   broker.Handler
 	quit      chan struct{}
 	done      chan struct{}
 	recs      chan kgo.FetchTopicPartition
+	kopts     broker.Options
+	partition int32
+	opts      broker.SubscribeOptions
+	handler   broker.Handler
+	connected *atomic.Uint32
 }
 
 type Subscriber struct {
-	c         *kgo.Client
-	topic     string
-	htracer   *hookTracer
-	opts      broker.SubscribeOptions
-	kopts     broker.Options
-	handler   broker.Handler
-	closed    bool
-	done      chan struct{}
 	consumers map[tp]*consumer
+	c         *kgo.Client
+	htracer   *hookTracer
+	topic     string
+
+	handler broker.Handler
+	done    chan struct{}
+	kopts   broker.Options
+	opts    broker.SubscribeOptions
+
+	connected *atomic.Uint32
 	sync.RWMutex
+	closed       bool
+	fatalOnError bool
 }
 
 func (s *Subscriber) Client() *kgo.Client {
@@ -63,21 +70,18 @@ func (s *Subscriber) Unsubscribe(ctx context.Context) error {
 	if s.closed {
 		return nil
 	}
-	select {
+
-	// case <-ctx.Done():
+	s.c.PauseFetchTopics(s.topic)
-	//	return ctx.Err()
+	s.c.CloseAllowingRebalance()
-	default:
+	kc := make(map[string][]int32)
-		s.c.PauseFetchTopics(s.topic)
+	for ctp := range s.consumers {
-		s.c.CloseAllowingRebalance()
+		kc[ctp.t] = append(kc[ctp.t], ctp.p)
-		kc := make(map[string][]int32)
-		for ctp := range s.consumers {
-			kc[ctp.t] = append(kc[ctp.t], ctp.p)
-		}
-		s.killConsumers(ctx, kc)
-		close(s.done)
-		s.closed = true
-		s.c.ResumeFetchTopics(s.topic)
 	}
+	s.killConsumers(ctx, kc)
+	close(s.done)
+	s.closed = true
+	s.c.ResumeFetchTopics(s.topic)
+
 	return nil
 }
 
@@ -141,8 +145,8 @@ func (s *Subscriber) poll(ctx context.Context) {
 			})
 
 			fetches.EachPartition(func(p kgo.FetchTopicPartition) {
-				tp := tp{p.Topic, p.Partition}
+				tps := tp{p.Topic, p.Partition}
-				s.consumers[tp].recs <- p
+				s.consumers[tps].recs <- p
 			})
 			s.c.AllowRebalance()
 		}
@@ -155,9 +159,9 @@ func (s *Subscriber) killConsumers(ctx context.Context, lost map[string][]int32)
 
 	for topic, partitions := range lost {
 		for _, partition := range partitions {
-			tp := tp{topic, partition}
+			tps := tp{topic, partition}
-			pc := s.consumers[tp]
+			pc := s.consumers[tps]
-			delete(s.consumers, tp)
+			delete(s.consumers, tps)
 			close(pc.quit)
 			if s.kopts.Logger.V(logger.DebugLevel) {
 				s.kopts.Logger.Debug(ctx, fmt.Sprintf("[kgo] waiting for work to finish topic %s partition %d", topic, partition))
@@ -168,11 +172,21 @@ func (s *Subscriber) killConsumers(ctx context.Context, lost map[string][]int32)
 	}
 }
 
+func (s *Subscriber) autocommit(_ *kgo.Client, _ *kmsg.OffsetCommitRequest, _ *kmsg.OffsetCommitResponse, err error) {
+	if err != nil {
+		s.connected.Store(0)
+		if s.fatalOnError {
+			s.kopts.Logger.Fatal(context.TODO(), "kgo.AutoCommitCallback error", err)
+		}
+	}
+}
+
 func (s *Subscriber) lost(ctx context.Context, _ *kgo.Client, lost map[string][]int32) {
-	if s.kopts.Logger.V(logger.DebugLevel) {
+	if s.kopts.Logger.V(logger.ErrorLevel) {
-		s.kopts.Logger.Debug(ctx, fmt.Sprintf("[kgo] lost %#+v", lost))
+		s.kopts.Logger.Error(ctx, fmt.Sprintf("[kgo] lost %#+v", lost))
 	}
 	s.killConsumers(ctx, lost)
+	// s.connected.Store(0)
 }
 
 func (s *Subscriber) revoked(ctx context.Context, c *kgo.Client, revoked map[string][]int32) {
@@ -182,6 +196,7 @@ func (s *Subscriber) revoked(ctx context.Context, c *kgo.Client, revoked map[str
 	s.killConsumers(ctx, revoked)
 	if err := c.CommitMarkedOffsets(ctx); err != nil {
 		s.kopts.Logger.Error(ctx, "[kgo] revoked CommitMarkedOffsets error", err)
+		// s.connected.Store(0)
 	}
 }
 
@@ -199,6 +214,7 @@ func (s *Subscriber) assigned(_ context.Context, c *kgo.Client, assigned map[str
 				handler:   s.handler,
 				kopts:     s.kopts,
 				opts:      s.opts,
+				connected: s.connected,
 			}
 			s.Lock()
 			s.consumers[tp{topic, partition}] = pc
@@ -260,6 +276,7 @@ func (pc *consumer) consume() {
 								pc.c.MarkCommitRecords(record)
 							} else {
 								eventPool.Put(p)
+								pc.connected.Store(0)
 								pc.kopts.Logger.Fatal(pc.kopts.Context, "[kgo] ErrLostMessage wtf?")
 								return
 							}
@@ -276,6 +293,7 @@ func (pc *consumer) consume() {
 						pc.kopts.Meter.Summary(semconv.SubscribeMessageLatencyMicroseconds, "endpoint", record.Topic, "topic", record.Topic).Update(te.Seconds())
 						pc.kopts.Meter.Histogram(semconv.SubscribeMessageDurationSeconds, "endpoint", record.Topic, "topic", record.Topic).Update(te.Seconds())
 						eventPool.Put(p)
+						pc.connected.Store(0)
 						pc.kopts.Logger.Fatal(pc.kopts.Context, "[kgo] Unmarshal err not handled wtf?")
 						sp.Finish()
 						return
@@ -313,6 +331,7 @@ func (pc *consumer) consume() {
 					pc.c.MarkCommitRecords(record)
 				} else {
 					eventPool.Put(p)
+					pc.connected.Store(0)
 					pc.kopts.Logger.Fatal(pc.kopts.Context, "[kgo] ErrLostMessage wtf?")
 					sp.SetStatus(tracer.SpanStatusError, "ErrLostMessage")
 					sp.Finish()

tracer.go

@@ -11,9 +11,9 @@ import (
 )
 
 type hookTracer struct {
+	tracer   tracer.Tracer
 	clientID string
 	group    string
-	tracer   tracer.Tracer
 }
 
 var messagingSystem = semconv.MessagingSystemKey.String("kafka")
@@ -68,7 +68,7 @@ func (m *hookTracer) OnProduceRecordBuffered(r *kgo.Record) {
 		r.Context, _ = m.tracer.Start(r.Context, "sdk.broker", opts...)
 	}
 
-	setHeaders(r, omd)
+	setHeaders(r, omd, metadata.HeaderContentType)
 }
 
 // OnProduceRecordUnbuffered continues and ends the "publish" span for an
@@ -135,7 +135,7 @@ func (m *hookTracer) OnFetchRecordBuffered(r *kgo.Record) {
 		r.Context, _ = m.tracer.Start(r.Context, "sdk.broker", opts...)
 	}
 
-	setHeaders(r, omd)
+	setHeaders(r, omd, metadata.HeaderContentType)
 }
 
 // OnFetchRecordUnbuffered continues and ends the "receive" span for an