Create v1.11.0-rc.1 (#4465)

* fix(prometheus.exporter.redis): pass MaxDistinctKeyGroups option to exporter (#4463)

Changes the `MaxDistinctKeyGroups` configuration option to be passed
into the Redis exporter instance. Previously it was accepted in Alloy's
configuration but not passed through, which would cause it to default
to zero regardless of how Alloy was configured.

Note that this would cause a cardinality increase in generated metrics
for instances of this component with key group metrics enabled.

Signed-off-by: stegosaurus21 <[email protected]>
Co-authored-by: Karl Persson <[email protected]>

* fix(loki.source.podlogs): enable proper log collection from Kubernetes Jobs and CronJobs (#4349)

* fix(loki.source.podlogs): enable proper log collection from Kubernetes Jobs and CronJobs

- Add job-aware termination logic with extended grace periods (10-60s) to ensure all job logs are captured
- Implement proper handling of pod deletion and race conditions between job completion and controller cleanup
- Separate concerns: containerTerminated() for regular pods, shouldStopTailingJobContainer() for jobs
- Enhance deduplication mechanisms to prevent duplicate log collection while ensuring comprehensive coverage
- Add comprehensive test coverage including unit tests and deduplication validation
- Fix race condition where job logs were missed for fast-completing or terminated jobs

This resolves the issue where the podlogs component would fail to scrape logs from short-lived or terminated Kubernetes jobs due to premature termination of log collection.

* rename container terminated

* refactor pod info extraction

* fix linting errors

* loki.source.journal: dont fail if journal files do not exist (#4462)

* Update changelog

* Upgrade Beyla to v2.6.5 (#4468)

---------

Signed-off-by: stegosaurus21 <[email protected]>
Co-authored-by: Jason Liu <[email protected]>
Co-authored-by: Quentin Bisson <[email protected]>
Co-authored-by: Piotr <[email protected]>
Co-authored-by: Rafael Roquetto <[email protected]>

Author: 5 people

CHANGELOG.md

@@ -10,7 +10,7 @@ internal API changes are not present.
 Main (unreleased)
 -----------------
 
-v1.11.0-rc.0
+v1.11.0-rc.1
 -----------------
 
 ### Breaking changes
@@ -176,6 +176,18 @@ v1.11.0-rc.0
 
 - Fix race conditions in `loki.source.syslog` where it could deadlock or cause port bind errors during config reload or shutdown. (@thampiotr)
 
+- **Fix `loki.source.podlogs` component to properly collect logs from Kubernetes Jobs and CronJobs.** Previously, the component would fail to scrape logs from short-lived or terminated jobs due to race conditions between job completion and pod discovery. The fix includes:
+  - Job-aware termination logic with extended grace periods (10-60 seconds) to ensure all logs are captured
+  - Proper handling of pod deletion and race conditions between job completion and controller cleanup
+  - Separation of concerns: `shouldStopTailingContainer()` handles standard Kubernetes restart policies for regular pods, while `shouldStopTailingJobContainer()` handles job-specific lifecycle with grace periods
+  - Enhanced deduplication mechanisms to prevent duplicate log collection while ensuring comprehensive coverage
+  - Comprehensive test coverage including unit tests and deduplication validation
+  This resolves the issue where job logs were being missed, particularly for fast-completing jobs or jobs that terminated before discovery. (@QuentinBisson)
+
+- Fix `prometheus.exporter.redis` component so that it no longer ignores the `MaxDistinctKeyGroups` configuration option. If key group metrics are enabled, this will increase the cardinality of the generated metrics. (@stegosaurus21)
+
+- Fix `loki.source.journal` creation failing with an error when the journal file is not found. (@thampiotr)
+
 v1.10.2
 -----------------
 

go.mod

@@ -63,7 +63,7 @@ require (
 	github.com/gorilla/mux v1.8.1
 	github.com/grafana/alloy-remote-config v0.0.11
 	github.com/grafana/alloy/syntax v0.1.0
-	github.com/grafana/beyla/v2 v2.6.4
+	github.com/grafana/beyla/v2 v2.6.5
 	github.com/grafana/catchpoint-prometheus-exporter v0.0.0-20250218151502-6e97feaee761
 	github.com/grafana/ckit v0.0.0-20250514165824-dd4adf36ad34
 	github.com/grafana/cloudflare-go v0.0.0-20230110200409-c627cf6792f2
@@ -1068,7 +1068,7 @@ exclude (
 	k8s.io/client-go v12.0.0+incompatible
 )
 
-replace go.opentelemetry.io/obi => github.com/grafana/opentelemetry-ebpf-instrumentation v1.2.4
+replace go.opentelemetry.io/obi => github.com/grafana/opentelemetry-ebpf-instrumentation v1.2.5
 
 replace go.opentelemetry.io/ebpf-profiler => github.com/grafana/opentelemetry-ebpf-profiler v0.0.202537-0.20250916114748-f2ff2fc6048c
 

go.sum

@@ -1124,8 +1124,8 @@ github.com/gosnmp/gosnmp v1.41.0 h1:6RI78g2ZsbLvpvJegcV98LapszRQnbvYNKSa5WbCll4=
 github.com/gosnmp/gosnmp v1.41.0/go.mod h1:CxVS6bXqmWZlafUj9pZUnQX5e4fAltqPcijxWpCitDo=
 github.com/grafana/alloy-remote-config v0.0.11 h1:IwYdQ38Tv/tEZuask8l29fdIcIzUP4HhY+VhMp9lqEk=
 github.com/grafana/alloy-remote-config v0.0.11/go.mod h1:kHE1usYo2WAVCikQkIXuoG1Clz8BSdiz3kF+DZSCQ4k=
-github.com/grafana/beyla/v2 v2.6.4 h1:nN4ljXuETUePYE4n8oOZjvRExaPW8cLjFTznD1hNl1E=
-github.com/grafana/beyla/v2 v2.6.4/go.mod h1:6X62WG9uxJedecIzmAYvK8iJXuM3xq1eRXOyz1fkAa4=
+github.com/grafana/beyla/v2 v2.6.5 h1:w2EiD8T8z2TS9dSG8axF4VGnmTN9qf9mrUoIdkH+efA=
+github.com/grafana/beyla/v2 v2.6.5/go.mod h1:kBxtZcdaDQ/a04ZNsbTii/uRkiQI3lTadhGbZ8gw1lo=
 github.com/grafana/cadvisor v0.0.0-20240729082359-1f04a91701e2 h1:ju6EcY2aEobeBg185ETtFCKj5WzaQ48qfkbsSRRQrF4=
 github.com/grafana/cadvisor v0.0.0-20240729082359-1f04a91701e2/go.mod h1:8sLW/G7rcFe1CKMaA4pYT4mX3P1xQVGqM6luzEzx/2g=
 github.com/grafana/catchpoint-prometheus-exporter v0.0.0-20250218151502-6e97feaee761 h1:dPJOIEwtQ8uR3Qa79pb/lsSFJQ6j4P9vpCUQ4fKimG4=
@@ -1166,8 +1166,8 @@ github.com/grafana/node_exporter v0.18.1-grafana-r01.0.20250806062222-612bdf9540
 github.com/grafana/node_exporter v0.18.1-grafana-r01.0.20250806062222-612bdf9540ec/go.mod h1:1qoZ74W1muNrTyeP2fUzUbg2UKD8AeX435xSCJNub7Y=
 github.com/grafana/opentelemetry-collector/featuregate v0.0.0-20240325174506-2fd1623b2ca0 h1:i/Ne0XwoRokYj52ZcSmnvuyID3h/uA91n0Ycg/grHU8=
 github.com/grafana/opentelemetry-collector/featuregate v0.0.0-20240325174506-2fd1623b2ca0/go.mod h1:mm8+xyQfgDmqhyegZRNIQmoKsNnDTwWKFLsdMoXAb7A=
-github.com/grafana/opentelemetry-ebpf-instrumentation v1.2.4 h1:2dIzC3i612KVMiG6A4sq53JS3YrCg5x4SXuP8zcLTrs=
-github.com/grafana/opentelemetry-ebpf-instrumentation v1.2.4/go.mod h1:EFQ1reX/fEsmFeyZn+G/lzMwGcEKlWBOIP8pECIvXZc=
+github.com/grafana/opentelemetry-ebpf-instrumentation v1.2.5 h1:PbcVxweZqvoIb58Yk0BFmCCeXsaqZI05XMHzZyQ8K7o=
+github.com/grafana/opentelemetry-ebpf-instrumentation v1.2.5/go.mod h1:EFQ1reX/fEsmFeyZn+G/lzMwGcEKlWBOIP8pECIvXZc=
 github.com/grafana/opentelemetry-ebpf-profiler v0.0.202537-0.20250916114748-f2ff2fc6048c h1:G+o7MkTXTgpEebJM5ctNMZw8DOPOatRM6lOLeqRPL50=
 github.com/grafana/opentelemetry-ebpf-profiler v0.0.202537-0.20250916114748-f2ff2fc6048c/go.mod h1:ajmdC82d8daScIWPT0Mmq95lvGmoBNMdAUBlWv/Hwbg=
 github.com/grafana/otel-profiling-go v0.5.1 h1:stVPKAFZSa7eGiqbYuG25VcqYksR6iWvF3YH66t4qL8=
@@ -1194,8 +1194,6 @@ github.com/grafana/tail v0.0.0-20230510142333-77b18831edf0 h1:bjh0PVYSVVFxzINqPF
 github.com/grafana/tail v0.0.0-20230510142333-77b18831edf0/go.mod h1:7t5XR+2IA8P2qggOAHTj/GCZfoLBle3OvNSYh1VkRBU=
 github.com/grafana/vmware_exporter v0.0.5-beta.0.20250218170317-73398ba08329 h1:Rs4H1yv2Abk3xE82qpyhMGGA8rswAOA0HQZde/BYkFo=
 github.com/grafana/vmware_exporter v0.0.5-beta.0.20250218170317-73398ba08329/go.mod h1:Z28219aViNlsLlPvuCnlgHDagRdZBAZ7JOnQg1b3eWg=
-github.com/grafana/walqueue v0.0.0-20250915204108-c3ca0631af46 h1:asmom2PA/hWPhzD6lx5+aK9HvqJnYzfB7NqmgVRuTgI=
-github.com/grafana/walqueue v0.0.0-20250915204108-c3ca0631af46/go.mod h1:LJm4P3SayTHSbHBYepsAf3WqlY/gwSYQyMs7OLLAi6A=
 github.com/grafana/walqueue v0.0.0-20250916201216-152b1f10cca2 h1:6HRYKHfWwdIoBF5jvVAYnARFHJiKe+++j1Oxh6A1mNw=
 github.com/grafana/walqueue v0.0.0-20250916201216-152b1f10cca2/go.mod h1:LJm4P3SayTHSbHBYepsAf3WqlY/gwSYQyMs7OLLAi6A=
 github.com/grobie/gomemcache v0.0.0-20230213081705-239240bbc445 h1:FlKQKUYPZ5yDCN248M3R7x8yu2E3yEZ0H7aLomE4EoE=

internal/component/loki/source/journal/journal.go

@@ -4,6 +4,7 @@ package journal
 
 import (
 	"context"
+	"fmt"
 	"os"
 	"path/filepath"
 	"sync"
@@ -12,13 +13,13 @@ import (
 	"github.com/grafana/loki/v3/clients/pkg/promtail/scrapeconfig"
 	"github.com/prometheus/common/model"
 
+	"github.com/grafana/alloy/internal/component"
 	"github.com/grafana/alloy/internal/component/common/loki"
 	"github.com/grafana/alloy/internal/component/common/loki/positions"
 	alloy_relabel "github.com/grafana/alloy/internal/component/common/relabel"
 	"github.com/grafana/alloy/internal/component/loki/source/journal/internal/target"
 	"github.com/grafana/alloy/internal/featuregate"
-
-	"github.com/grafana/alloy/internal/component"
+	"github.com/grafana/alloy/internal/runtime/logging/level"
 )
 
 func init() {
@@ -37,13 +38,16 @@ var _ component.Component = (*Component)(nil)
 
 // Component represents reading from a journal
 type Component struct {
-	mut       sync.RWMutex
-	t         *target.JournalTarget
-	metrics   *target.Metrics
-	o         component.Options
-	handler   chan loki.Entry
-	positions positions.Positions
-	receivers []loki.LogsReceiver
+	mut           sync.RWMutex
+	t             *target.JournalTarget
+	metrics       *target.Metrics
+	o             component.Options
+	handler       chan loki.Entry
+	positions     positions.Positions
+	receivers     []loki.LogsReceiver
+	argsUpdated   chan struct{}
+	args          Arguments
+	healthErr     error
 }
 
 // New creates a new  component.
@@ -69,11 +73,13 @@ func New(o component.Options, args Arguments) (*Component, error) {
 	}
 
 	c := &Component{
-		metrics:   target.NewMetrics(o.Registerer),
-		o:         o,
-		handler:   make(chan loki.Entry),
-		positions: positionsFile,
-		receivers: args.Receivers,
+		metrics:       target.NewMetrics(o.Registerer),
+		o:             o,
+		handler:       make(chan loki.Entry),
+		positions:     positionsFile,
+		receivers:     args.Receivers,
+		argsUpdated:   make(chan struct{}, 1),
+		args:          args,
 	}
 	err = c.Update(args)
 	return c, err
@@ -84,7 +90,10 @@ func (c *Component) Run(ctx context.Context) error {
 	defer func() {
 		c.mut.RLock()
 		if c.t != nil {
-			c.t.Stop()
+			err := c.t.Stop()
+			if err != nil {
+				level.Warn(c.o.Logger).Log("msg", "error stopping journal target", "err", err)
+			}
 		}
 		c.mut.RUnlock()
 
@@ -103,6 +112,27 @@ func (c *Component) Run(ctx context.Context) error {
 				r.Chan() <- lokiEntry
 			}
 			c.mut.RUnlock()
+		case <-c.argsUpdated:
+			c.mut.Lock()
+			if c.t != nil {
+				err := c.t.Stop()
+				if err != nil {
+					level.Error(c.o.Logger).Log("msg", "error stopping journal target", "err", err)
+				}
+				c.t = nil
+			}
+			rcs := alloy_relabel.ComponentToPromRelabelConfigs(c.args.RelabelRules)
+			entryHandler := loki.NewEntryHandler(c.handler, func() {})
+
+			newTarget, err := target.NewJournalTarget(c.metrics, c.o.Logger, entryHandler, c.positions, c.o.ID, rcs, convertArgs(c.o.ID, c.args))
+			if err != nil {
+				level.Error(c.o.Logger).Log("msg", "error creating journal target", "err", err, "path", c.args.Path)
+				c.healthErr = fmt.Errorf("error creating journal target: %w", err)
+			} else {
+				c.t = newTarget
+				c.healthErr = nil
+			}
+			c.mut.Unlock()
 		}
 	}
 }
@@ -112,21 +142,31 @@ func (c *Component) Update(args component.Arguments) error {
 	newArgs := args.(Arguments)
 	c.mut.Lock()
 	defer c.mut.Unlock()
-	if c.t != nil {
-		err := c.t.Stop()
-		if err != nil {
-			return err
-		}
+	c.args = newArgs
+	select {
+	case c.argsUpdated <- struct{}{}:
+	default: // Update notification already sent
 	}
-	rcs := alloy_relabel.ComponentToPromRelabelConfigs(newArgs.RelabelRules)
-	entryHandler := loki.NewEntryHandler(c.handler, func() {})
+	return nil
+}
 
-	newTarget, err := target.NewJournalTarget(c.metrics, c.o.Logger, entryHandler, c.positions, c.o.ID, rcs, convertArgs(c.o.ID, newArgs))
-	if err != nil {
-		return err
+// CurrentHealth implements component.HealthComponent. It returns an unhealthy
+// status if the server has terminated.
+func (c *Component) CurrentHealth() component.Health {
+	c.mut.RLock()
+	defer c.mut.RUnlock()
+	if c.healthErr == nil {
+		return component.Health{
+			Health:     component.HealthTypeHealthy,
+			Message:    "journal target is running",
+			UpdateTime: time.Now(),
+		}
+	}
+	return component.Health{
+		Health:     component.HealthTypeUnhealthy,
+		Message:    c.healthErr.Error(),
+		UpdateTime: time.Now(),
 	}
-	c.t = newTarget
-	return nil
 }
 
 func convertArgs(job string, a Arguments) *scrapeconfig.JournalTargetConfig {