prometheus best practices, TLS and basic auth support

.github/workflows/exporter-ci.yml

@@ -24,13 +24,13 @@ jobs:
   build:
     runs-on: ubuntu-latest
     steps:
-    - uses: actions/setup-go@v2
-      with:
-        go-version: ^1.14
-      id: go
-    - uses: actions/checkout@v2
+    - uses: actions/checkout@v3
       with:
         fetch-depth: 0
+    - uses: actions/setup-go@v3
+      with:
+        go-version-file: 'go.mod'
+      id: go
     - name: static analysis
       run: make static-checks
     - name: test
@@ -43,7 +43,7 @@ jobs:
         for FILE in build/bin/*; do
           gzip $FILE
         done
-    - uses: actions/upload-artifact@v2
+    - uses: actions/upload-artifact@v3
       with:
         name: ha_cluster_exporter
         path: build/bin
@@ -104,10 +104,10 @@ jobs:
     if: github.event.release
     runs-on: ubuntu-latest
     steps:
-    - uses: actions/download-artifact@v2
+    - uses: actions/download-artifact@v3
       with:
         name: ha_cluster_exporter
-    - uses: AButler/[email protected]
+    - uses: softprops/action-gh-release@v1
       with:
         files: 'ha_cluster_exporter-*'
-        repo-token: ${{ secrets.GITHUB_TOKEN }}
+        token: ${{ secrets.GITHUB_TOKEN }}

.gitignore

@@ -17,3 +17,6 @@ ha_cluster_exporter
 # we don't use vendoring
 /vendor
 /build
+
+# ignore tmp promu config
+.promu.release.yml

.promu.yml

@@ -0,0 +1,21 @@
+go:
+    version: 1.16
+    cgo: false
+repository:
+    path: github.com/ClusterLabs/ha_cluster_exporter
+build:
+    flags: -a -tags netgo
+    ldflags: |
+        -X github.com/prometheus/common/version.Version={{.Version}}
+        -X github.com/prometheus/common/version.Revision={{.Revision}}
+        -X github.com/prometheus/common/version.Branch={{.Branch}}
+        -X github.com/prometheus/common/version.BuildUser={{user}}@{{host}}
+        -X github.com/prometheus/common/version.BuildDate={{date "20060102-15:04:05"}}
+    binaries:
+        - name: ha_cluster_exporter-amd64
+        - name: ha_cluster_exporter-arm64
+        - name: ha_cluster_exporter-ppc64le
+        - name: ha_cluster_exporter-s390x
+tarball:
+    files:
+        - LICENSE

Makefile

@@ -1,9 +1,20 @@
+GO    := GO111MODULE=on go
+FIRST_GOPATH := $(firstword $(subst :, ,$(shell $(GO) env GOPATH)))
+GOHOSTOS     ?= $(shell $(GO) env GOHOSTOS)
+GOHOSTARCH   ?= $(shell $(GO) env GOHOSTARCH)
+ifeq (arm, $(GOHOSTARCH))
+	GOHOSTARM ?= $(shell GOARM= $(GO) env GOARM)
+	GO_BUILD_PLATFORM ?= $(GOHOSTOS)-$(GOHOSTARCH)v$(GOHOSTARM)
+else
+	GO_BUILD_PLATFORM ?= $(GOHOSTOS)-$(GOHOSTARCH)
+endif
+PROMU        := $(FIRST_GOPATH)/bin/promu
+PROMU_VERSION ?= 0.13.0
+PROMU_URL     := https://github.com/prometheus/promu/releases/download/v$(PROMU_VERSION)/promu-$(PROMU_VERSION).$(GO_BUILD_PLATFORM).tar.gz
+
 # this is the what ends up in the RPM "Version" field and embedded in the --version CLI flag
 VERSION ?= $(shell .ci/get_version_from_git.sh)
 
-# this will be used as the build date by the Go compile task
-DATE = $(shell date --iso-8601=seconds)
-
 # if you want to release to OBS, this must be a remotely available Git reference
 REVISION ?= $(shell git rev-parse --abbrev-ref HEAD)
 
@@ -19,47 +30,65 @@ ARCHS ?= amd64 arm64 ppc64le s390x
 
 default: clean mod-tidy generate fmt vet-check test build
 
-build: amd64
+promu-prepare: 
+	sed "s/{{.Version}}/$(VERSION)/" .promu.yml >.promu.release.yml
+	mkdir -p build/bin
+
+# from https://github.com/prometheus/prometheus/blob/main/Makefile.common
+$(PROMU):
+	$(eval PROMU_TMP := $(shell mktemp -d))
+	curl -s -L $(PROMU_URL) | tar -xvzf - -C $(PROMU_TMP)
+	mkdir -p $(FIRST_GOPATH)/bin
+	cp $(PROMU_TMP)/promu-$(PROMU_VERSION).$(GO_BUILD_PLATFORM)/promu $(FIRST_GOPATH)/bin/promu
+	rm -r $(PROMU_TMP)
+
+build:
+	$(MAKE) clean
+	$(MAKE) promu-prepare $(PROMU)
+	$(MAKE) amd64
 
-build-all: clean $(ARCHS)
+build-all:
+	$(MAKE) clean
+	$(MAKE) promu-prepare $(PROMU)
+	$(MAKE) $(ARCHS)
 
 $(ARCHS):
-	@mkdir -p build/bin
-	CGO_ENABLED=0 GOOS=linux GOARCH=$@ go build -trimpath -ldflags "-s -w -X main.version=$(VERSION) -X main.buildDate=$(DATE)" -o build/bin/ha_cluster_exporter-$@
+	GOOS=linux GOARCH=$@ $(PROMU) build --config .promu.release.yml --prefix=build/bin ha_cluster_exporter-$@
 
 install:
-	go install
+	$(GO) install
 
 static-checks: vet-check fmt-check
 
 vet-check:
-	go vet ./...
+	$(GO) vet ./...
 
 fmt:
-	go fmt ./...
+	$(GO) fmt ./...
 
 mod-tidy:
-	go mod tidy
+	$(GO) mod tidy
 
 fmt-check:
 	.ci/go_lint.sh
 
 generate:
-	go generate ./...
+	$(GO) generate ./...
 
 test:
-	go test -v ./...
+	$(GO) test -v ./...
 
 checks: static-checks test
 
 coverage:
 	@mkdir -p build
-	go test -cover -coverprofile=build/coverage ./...
-	go tool cover -html=build/coverage
+	$(GO) test -cover -coverprofile=build/coverage ./...
+	$(GO) tool cover -html=build/coverage
 
 clean:
-	go clean
+	$(GO) clean
 	rm -rf build
+	rm -f .promu.release.yml
 
 exporter-obs-workdir: build/obs/prometheus-ha_cluster_exporter
 build/obs/prometheus-ha_cluster_exporter:

README.md

@@ -98,6 +98,47 @@ The first match has precedence, and the CLI flags have precedence over the confi
 
 Please refer to the example [YAML configuration](ha_cluster_exporter.yaml) for more details.
 
+Additional CLI flags can also be passed via `/etc/sysconfig/prometheus-ha_cluster_exporter`.
+
+#### General Flags
+
+Name                                       | Description
+----                                       | -----------
+web.listen-address                         | Address to listen on for web interface and telemetry (default `:9664`).
+web.telemetry-path                         | Path under which to expose metrics (default `/metrics`).
+web.config.file                            | Path to a [web configuration file](#tls-and-basic-authentication) (default `/etc/ha_cluster_exporter.web.yaml`).
+log.level                                  | Logging verbosity (default `info`).
+version                                    | Print the version information.
+
+##### Deprecated Flags
+Name                                       | Description
+----                                       | -----------
+address                                    | deprecated: please use --web.listen-address or --web.config.file to use Prometheus Exporter Toolkit
+port                                       | deprecated: please use --web.listen-address or --web.config.file to use Prometheus Exporter Toolkit
+log-level                                  | deprecated: please use log.level
+enable-timestamps                          | deprecated: server-side metric timestamping is discouraged by Prometheus best-practices and should be avoided
+
+#### Collector Flags
+
+Name                                       | Description
+----                                       | -----------
+crm-mon-path                               | Path to crm_mon executable (default `/usr/sbin/crm_mon`).
+cibadmin-path                              | Path to cibadmin executable (default `/usr/sbin/cibadmin`).
+corosync-cfgtoolpath-path                  | Path to corosync-cfgtool executable (default `/usr/sbin/corosync-cfgtool`).
+corosync-quorumtool-path                   | Path to corosync-quorumtool executable (default `/usr/sbin/corosync-quorumtool`).
+sbd-path                                   | Path to sbd executable (default `/usr/sbin/sbd`).
+sbd-config-path                            | Path to sbd configuration (default `/etc/sysconfig/sbd`).
+drbdsetup-path                             | Path to drbdsetup executable (default `/sbin/drbdsetup`).
+drbdsplitbrain-path                        | Path to drbd splitbrain hooks temporary files (default `/var/run/drbd/splitbrain`).
+
+### TLS and basic authentication
+
+The ha_cluster_exporter supports TLS and basic authentication.
+
+To use TLS and/or basic authentication, you need to pass a configuration file
+using the `--web.config.file` parameter. The format of the file is described
+[in the exporter-toolkit repository](https://github.com/prometheus/exporter-toolkit/blob/master/docs/web-configuration.md).
+
 ### systemd integration
 
 A [systemd unit file](ha_cluster_exporter.service) is provided with the RPM packages. You can enable and start it as usual:  
@@ -114,7 +155,7 @@ We recommend having a look at the [design document](doc/design.md) and the [deve
 
 ## License
 
-Copyright 2019-2020 SUSE LLC
+Copyright 2019-2022 SUSE LLC
 
 Licensed under the Apache License, Version 2.0 (the "License");
 you may not use this file except in compliance with the License.

collector/corosync/corosync.go

@@ -3,23 +3,24 @@ package corosync
 import (
 	"os/exec"
 
+	"github.com/go-kit/log"
+	"github.com/go-kit/log/level"
 	"github.com/pkg/errors"
 	"github.com/prometheus/client_golang/prometheus"
-	log "github.com/sirupsen/logrus"
 
 	"github.com/ClusterLabs/ha_cluster_exporter/collector"
 )
 
 const subsystem = "corosync"
 
-func NewCollector(cfgToolPath string, quorumToolPath string) (*corosyncCollector, error) {
+func NewCollector(cfgToolPath string, quorumToolPath string, timestamps bool, logger log.Logger) (*corosyncCollector, error) {
 	err := collector.CheckExecutables(cfgToolPath, quorumToolPath)
 	if err != nil {
 		return nil, errors.Wrapf(err, "could not initialize '%s' collector", subsystem)
 	}
 
 	c := &corosyncCollector{
-		collector.NewDefaultCollector(subsystem),
+		collector.NewDefaultCollector(subsystem, timestamps, logger),
 		cfgToolPath,
 		quorumToolPath,
 		NewParser(),
@@ -41,7 +42,7 @@ type corosyncCollector struct {
 }
 
 func (c *corosyncCollector) CollectWithError(ch chan<- prometheus.Metric) error {
-	log.Debugln("Collecting corosync metrics...")
+	level.Debug(c.Logger).Log("msg", "Collecting corosync metrics...")
 
 	// We suppress the exec errors because if any interface is faulty the tools will exit with code 1, but we still want to parse the output.
 	cfgToolOutput, _ := exec.Command(c.cfgToolPath, "-s").Output()
@@ -62,9 +63,11 @@ func (c *corosyncCollector) CollectWithError(ch chan<- prometheus.Metric) error
 }
 
 func (c *corosyncCollector) Collect(ch chan<- prometheus.Metric) {
+	level.Debug(c.Logger).Log("msg", "Collecting corosync metrics...")
+
 	err := c.CollectWithError(ch)
 	if err != nil {
-		log.Warnf("'%s' collector scrape failed: %s", c.GetSubsystem(), err)
+		level.Warn(c.Logger).Log("msg", c.GetSubsystem()+" collector scrape failed", "err", err)
 	}
 }
 

collector/corosync/corosync_test.go

@@ -3,45 +3,46 @@ package corosync
 import (
 	"testing"
 
+	"github.com/go-kit/log"
 	"github.com/stretchr/testify/assert"
 
 	assertcustom "github.com/ClusterLabs/ha_cluster_exporter/internal/assert"
 )
 
 func TestNewCorosyncCollector(t *testing.T) {
-	_, err := NewCollector("../../test/fake_corosync-cfgtool.sh", "../../test/fake_corosync-quorumtool.sh")
+	_, err := NewCollector("../../test/fake_corosync-cfgtool.sh", "../../test/fake_corosync-quorumtool.sh", false, log.NewNopLogger())
 	assert.Nil(t, err)
 }
 
 func TestNewCorosyncCollectorChecksCfgtoolExistence(t *testing.T) {
-	_, err := NewCollector("../../test/nonexistent", "../../test/fake_corosync-quorumtool.sh")
+	_, err := NewCollector("../../test/nonexistent", "../../test/fake_corosync-quorumtool.sh", false, log.NewNopLogger())
 
 	assert.Error(t, err)
 	assert.Contains(t, err.Error(), "'../../test/nonexistent' does not exist")
 }
 
 func TestNewCorosyncCollectorChecksQuorumtoolExistence(t *testing.T) {
-	_, err := NewCollector("../../test/fake_corosync-cfgtool.sh", "../../test/nonexistent")
+	_, err := NewCollector("../../test/fake_corosync-cfgtool.sh", "../../test/nonexistent", false, log.NewNopLogger())
 
 	assert.Error(t, err)
 	assert.Contains(t, err.Error(), "'../../test/nonexistent' does not exist")
 }
 
 func TestNewCorosyncCollectorChecksCfgtoolExecutableBits(t *testing.T) {
-	_, err := NewCollector("../../test/dummy", "../../test/fake_corosync-quorumtool.sh")
+	_, err := NewCollector("../../test/dummy", "../../test/fake_corosync-quorumtool.sh", false, log.NewNopLogger())
 
 	assert.Error(t, err)
 	assert.Contains(t, err.Error(), "'../../test/dummy' is not executable")
 }
 
 func TestNewCorosyncCollectorChecksQuorumtoolExecutableBits(t *testing.T) {
-	_, err := NewCollector("../../test/fake_corosync-cfgtool.sh", "../../test/dummy")
+	_, err := NewCollector("../../test/fake_corosync-cfgtool.sh", "../../test/dummy", false, log.NewNopLogger())
 
 	assert.Error(t, err)
 	assert.Contains(t, err.Error(), "'../../test/dummy' is not executable")
 }
 
 func TestCorosyncCollector(t *testing.T) {
-	collector, _ := NewCollector("../../test/fake_corosync-cfgtool.sh", "../../test/fake_corosync-quorumtool.sh")
+	collector, _ := NewCollector("../../test/fake_corosync-cfgtool.sh", "../../test/fake_corosync-quorumtool.sh", false, log.NewNopLogger())
 	assertcustom.Metrics(t, collector, "corosync.metrics")
 }

collector/default_collector.go

@@ -2,9 +2,9 @@ package collector
 
 import (
 	"github.com/ClusterLabs/ha_cluster_exporter/internal/clock"
+	"github.com/go-kit/log"
 	"github.com/pkg/errors"
 	"github.com/prometheus/client_golang/prometheus"
-	config "github.com/spf13/viper"
 	"os"
 )
 
@@ -18,13 +18,17 @@ type DefaultCollector struct {
 	subsystem   string
 	descriptors map[string]*prometheus.Desc
 	Clock       clock.Clock
+	timestamps  bool
+	Logger      log.Logger
 }
 
-func NewDefaultCollector(subsystem string) DefaultCollector {
+func NewDefaultCollector(subsystem string, timestamps bool, logger log.Logger) DefaultCollector {
 	return DefaultCollector{
 		subsystem,
 		make(map[string]*prometheus.Desc),
 		&clock.SystemClock{},
+		timestamps,
+		logger,
 	}
 }
 
@@ -67,7 +71,7 @@ func (c *DefaultCollector) MakeCounterMetric(name string, value float64, labelVa
 func (c *DefaultCollector) makeMetric(name string, value float64, valueType prometheus.ValueType, labelValues ...string) prometheus.Metric {
 	desc := c.GetDescriptor(name)
 	metric := prometheus.MustNewConstMetric(desc, valueType, value, labelValues...)
-	if config.GetBool("enable-timestamps") {
+	if c.timestamps == true {
 		metric = prometheus.NewMetricWithTimestamp(c.Clock.Now(), metric)
 	}
 	return metric