use errgroup instead of WaitGroup for concurrent snapshot builders

Replace sync.WaitGroup with errgroup.Group in buildSubSnapshots so errors from
sub-builders (buildSyscallSnapshots, buildFileSnapshots, buildProcessSnapshots,
buildHistogramSnapshot) are captured and propagated rather than silently dropped.

Change Engine.Snapshot() to return (*Snapshot, error), update runtime.SnapshotSource
and dashboard.SnapshotSource interfaces accordingly, and adjust all callers in
tui.go, dashboard/model.go, and the test helpers.

Each sub-builder now returns (result, error); the error return is currently
always nil but establishes the contract for future validation. The per-type
Snapshot() convenience methods (histogram, syscall, file, process) panic on
error since they are internal helpers where failure would be a programming bug.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

3 files changed, 31 insertions, 12 deletions

diff --git a/internal/tui/dashboard/model.go b/internal/tui/dashboard/model.go
index 8a3c5d4..cc2a052 100644
--- a/internal/tui/dashboard/model.go
+++ b/internal/tui/dashboard/model.go
@@ -874,11 +874,18 @@ func (m Model) maxProcessesRows() int {
 	return m.snapshotOrZero().ProcessesCount()
 }
 
+// snapshot returns the latest engine snapshot, or nil if the engine is nil or
+// returns an error. Errors are silently dropped here because the dashboard
+// renders the last successful snapshot on transient failures.
 func (m Model) snapshot() *statsengine.Snapshot {
 	if m.engine == nil {
 		return nil
 	}
-	return m.engine.Snapshot()
+	snap, err := m.engine.Snapshot()
+	if err != nil {
+		return nil
+	}
+	return snap
 }
 
 func (m Model) snapshotOrZero() statsengine.Snapshot {
@@ -893,10 +900,12 @@ func (m *Model) resetBaselineCmd() tea.Cmd {
 		m.liveTrie.Reset()
 	}
 
+	// Errors from Snapshot are silently dropped here; the dashboard will
+	// continue to display the last successful snapshot.
 	var snap *statsengine.Snapshot
 	if resettable, ok := m.engine.(resettableSnapshotSource); ok {
 		resettable.Reset()
-		snap = resettable.Snapshot()
+		snap, _ = resettable.Snapshot()
 	} else {
 		snap = m.snapshot()
 	}
diff --git a/internal/tui/tui.go b/internal/tui/tui.go
index eee4252..2cebaf2 100644
--- a/internal/tui/tui.go
+++ b/internal/tui/tui.go
@@ -225,6 +225,9 @@ func (r *runtimeBindings) advanceFilterEpoch() uint64 {
 	return r.filterEpoch.Add(1)
 }
 
+// resetDashboardSnapshotSource resets the dashboard snapshot source if it
+// implements the resettable interface, then returns a fresh snapshot. Errors
+// from Snapshot are silently dropped since callers handle a nil snapshot.
 func (r *runtimeBindings) resetDashboardSnapshotSource() *statsengine.Snapshot {
 	src := r.dashboardSnapshotSource()
 	if src == nil {
@@ -232,10 +235,11 @@ func (r *runtimeBindings) resetDashboardSnapshotSource() *statsengine.Snapshot {
 	}
 	if resettable, ok := src.(interface {
 		Reset()
-		Snapshot() *statsengine.Snapshot
+		Snapshot() (*statsengine.Snapshot, error)
 	}); ok {
 		resettable.Reset()
-		return resettable.Snapshot()
+		snap, _ := resettable.Snapshot()
+		return snap
 	}
 	return nil
 }
@@ -1216,13 +1220,16 @@ type lateBoundDashboardSource struct {
 	runtime *runtimeBindings
 }
 
-func (s lateBoundDashboardSource) Snapshot() *statsengine.Snapshot {
+// Snapshot returns a point-in-time dashboard snapshot from the underlying
+// source, or (nil, nil) when no source is available. Errors are forwarded to
+// the caller so they can decide how to handle a failed snapshot build.
+func (s lateBoundDashboardSource) Snapshot() (*statsengine.Snapshot, error) {
 	if s.runtime == nil {
-		return nil
+		return nil, nil
 	}
 	source := s.runtime.dashboardSnapshotSource()
 	if source == nil {
-		return nil
+		return nil, nil
 	}
 	return source.Snapshot()
 }
diff --git a/internal/tui/tui_test.go b/internal/tui/tui_test.go
index 5f007e4..d76ccde 100644
--- a/internal/tui/tui_test.go
+++ b/internal/tui/tui_test.go
@@ -459,8 +459,8 @@ type fakeDashboardSource struct {
 	snap *statsengine.Snapshot
 }
 
-func (f fakeDashboardSource) Snapshot() *statsengine.Snapshot {
-	return f.snap
+func (f fakeDashboardSource) Snapshot() (*statsengine.Snapshot, error) {
+	return f.snap, nil
 }
 
 type fakeResettableDashboardSource struct {
@@ -468,8 +468,8 @@ type fakeResettableDashboardSource struct {
 	resetCalls int
 }
 
-func (f *fakeResettableDashboardSource) Snapshot() *statsengine.Snapshot {
-	return f.snap
+func (f *fakeResettableDashboardSource) Snapshot() (*statsengine.Snapshot, error) {
+	return f.snap, nil
 }
 
 func (f *fakeResettableDashboardSource) Reset() {
@@ -484,7 +484,10 @@ func TestDashboardRefreshPicksLateBoundSource(t *testing.T) {
 	want := &statsengine.Snapshot{TotalSyscalls: 77}
 	runtime.SetDashboardSnapshotSource(fakeDashboardSource{snap: want})
 
-	got := source.Snapshot()
+	got, err := source.Snapshot()
+	if err != nil {
+		t.Fatalf("unexpected snapshot error: %v", err)
+	}
 	if got != want {
 		t.Fatalf("expected late-bound source to use latest runtime source")
 	}