Set slogger on rungroup after logshipping is set up, so that rungroup…

… logs will be shipped

cmd/launcher/desktop.go

@@ -111,7 +111,8 @@ func runDesktop(_ *multislogger.MultiSlogger, args []string) error {
 		)
 	}
 
-	runGroup := rungroup.NewRunGroup(slogger)
+	runGroup := rungroup.NewRunGroup()
+	runGroup.SetSlogger(slogger)
 
 	// listen for signals
 	runGroup.Add("desktopSignalListener", func() error {

cmd/launcher/launcher.go

@@ -206,7 +206,7 @@ func runLauncher(ctx context.Context, cancel func(), multiSlogger, systemMultiSl
 	}
 
 	// create a rungroup for all the actors we create to allow for easy start/stop
-	runGroup := rungroup.NewRunGroup(slogger)
+	runGroup := rungroup.NewRunGroup()
 
 	// Need to set up the log shipper so that we can get the logger early
 	// and pass it to the various systems.
@@ -246,6 +246,10 @@ func runLauncher(ctx context.Context, cancel func(), multiSlogger, systemMultiSl
 		startupSpan.AddEvent("log_shipper_init_completed")
 	}
 
+	// Now that log shipping is set up, set the slogger on the rungroup so that rungroup logs
+	// will also be shipped.
+	runGroup.SetSlogger(k.Slogger())
+
 	startupSettingsWriter, err := startupsettings.OpenWriter(ctx, k)
 	if err != nil {
 		return fmt.Errorf("creating startup db: %w", err)

pkg/rungroup/rungroup.go

@@ -12,6 +12,7 @@ import (
 	"time"
 
 	"github.com/kolide/launcher/ee/gowrapper"
+	"github.com/kolide/launcher/pkg/log/multislogger"
 	"golang.org/x/sync/semaphore"
 )
 
@@ -38,9 +39,9 @@ const (
 	executeReturnTimeout = 5 * time.Second  // After interrupted, how long for all actors to exit their `execute` functions
 )
 
-func NewRunGroup(slogger *slog.Logger) *Group {
+func NewRunGroup() *Group {
 	return &Group{
-		slogger: slogger.With("component", "run_group"),
+		slogger: multislogger.NewNopLogger(),
 		actors:  make([]rungroupActor, 0),
 	}
 }
@@ -49,6 +50,10 @@ func (g *Group) Add(name string, execute func() error, interrupt func(error)) {
 	g.actors = append(g.actors, rungroupActor{name, execute, interrupt})
 }
 
+func (g *Group) SetSlogger(slogger *slog.Logger) {
+	g.slogger = slogger.With("component", "run_group")
+}
+
 func (g *Group) Run() error {
 	if len(g.actors) == 0 {
 		return nil

pkg/rungroup/rungroup_test.go

@@ -7,22 +7,21 @@ import (
 	"testing"
 	"time"
 
-	"github.com/kolide/launcher/pkg/log/multislogger"
 	"github.com/kolide/launcher/pkg/threadsafebuffer"
 	"github.com/stretchr/testify/require"
 )
 
 func TestRun_NoActors(t *testing.T) {
 	t.Parallel()
 
-	testRunGroup := NewRunGroup(multislogger.NewNopLogger())
+	testRunGroup := NewRunGroup()
 	require.NoError(t, testRunGroup.Run())
 }
 
 func TestRun_MultipleActors(t *testing.T) {
 	t.Parallel()
 
-	testRunGroup := NewRunGroup(multislogger.NewNopLogger())
+	testRunGroup := NewRunGroup()
 
 	groupReceivedInterrupts := make(chan struct{}, 3)
 
@@ -93,7 +92,7 @@ func TestRun_MultipleActors(t *testing.T) {
 func TestRun_MultipleActors_InterruptTimeout(t *testing.T) {
 	t.Parallel()
 
-	testRunGroup := NewRunGroup(multislogger.NewNopLogger())
+	testRunGroup := NewRunGroup()
 
 	groupReceivedInterrupts := make(chan struct{}, 3)
 
@@ -166,7 +165,7 @@ func TestRun_MultipleActors_InterruptTimeout(t *testing.T) {
 func TestRun_MultipleActors_ExecuteReturnTimeout(t *testing.T) {
 	t.Parallel()
 
-	testRunGroup := NewRunGroup(multislogger.NewNopLogger())
+	testRunGroup := NewRunGroup()
 
 	groupReceivedInterrupts := make(chan struct{}, 3)
 	// Keep track of when `execute`s return so we give testRunGroup.Run enough time to do its thing
@@ -250,7 +249,8 @@ func TestRun_RecoversAndLogsPanic(t *testing.T) {
 		Level: slog.LevelDebug,
 	}))
 
-	testRunGroup := NewRunGroup(slogger)
+	testRunGroup := NewRunGroup()
+	testRunGroup.SetSlogger(slogger)
 
 	// Actor that will panic in its execute function
 	testRunGroup.Add("panickingActor", func() error {