Fix review issues: detached contexts, loop device leak, timer leak, size_bytes

- Use context.Background() with timeout in destroySandbox/failBuild so
  cleanup and DB writes survive parent context cancellation on shutdown
- Fix loop device refcount leak in FlattenRootfs when dmDevice is nil
- Replace time.After with time.NewTimer in healthcheck polling to avoid
  goroutine leak when healthcheck passes early
- Capture size_bytes from CreateSnapshot/FlattenRootfs RPC responses
  instead of hardcoding 0 in the templates table insert
- Avoid leaking internal error details to API clients in build handler

internal/api/handlers_builds.go

@@ -3,6 +3,7 @@ package api
 import (
 	"encoding/json"
 	"fmt"
+	"log/slog"
 	"net/http"
 	"time"
 
@@ -119,7 +120,8 @@ func (h *buildHandler) Create(w http.ResponseWriter, r *http.Request) {
 		MemoryMB:     req.MemoryMB,
 	})
 	if err != nil {
-		writeError(w, http.StatusInternalServerError, "build_error", err.Error())
+		slog.Error("failed to create build", "error", err)
+		writeError(w, http.StatusInternalServerError, "build_error", "failed to create build")
 		return
 	}
 

internal/sandbox/manager.go

@@ -839,6 +839,8 @@ func (m *Manager) FlattenRootfs(ctx context.Context, sandboxID, name string) (in
 
 	outputPath := snapshot.RootfsPath(m.cfg.ImagesDir, name)
 	if sb.dmDevice == nil {
+		m.cleanupDM(sb)
+		warnErr("template dir cleanup error", name, snapshot.Remove(m.cfg.ImagesDir, name))
 		return 0, fmt.Errorf("sandbox %s has no dm device", sandboxID)
 	}
 

internal/service/build.go

@@ -255,6 +255,7 @@ func (s *BuildService) executeBuild(ctx context.Context, buildID string) {
 	}
 
 	// Healthcheck or direct snapshot.
+	var sizeBytes int64
 	if build.Healthcheck.Valid && build.Healthcheck.String != "" {
 		log.Info("running healthcheck", "cmd", build.Healthcheck.String)
 		if err := s.waitForHealthcheck(ctx, agent, sandboxID, build.Healthcheck.String); err != nil {
@@ -265,25 +266,29 @@ func (s *BuildService) executeBuild(ctx context.Context, buildID string) {
 
 		// Healthcheck passed → full snapshot (with memory/CPU state).
 		log.Info("healthcheck passed, creating snapshot")
-		if _, err := agent.CreateSnapshot(ctx, connect.NewRequest(&pb.CreateSnapshotRequest{
+		snapResp, err := agent.CreateSnapshot(ctx, connect.NewRequest(&pb.CreateSnapshotRequest{
 			SandboxId: sandboxID,
 			Name:      build.Name,
-		})); err != nil {
+		}))
+		if err != nil {
 			s.destroySandbox(ctx, agent, sandboxID)
 			s.failBuild(ctx, buildID, fmt.Sprintf("create snapshot failed: %v", err))
 			return
 		}
+		sizeBytes = snapResp.Msg.SizeBytes
 	} else {
 		// No healthcheck → image-only template (rootfs only).
 		log.Info("no healthcheck, flattening rootfs")
-		if _, err := agent.FlattenRootfs(ctx, connect.NewRequest(&pb.FlattenRootfsRequest{
+		flatResp, err := agent.FlattenRootfs(ctx, connect.NewRequest(&pb.FlattenRootfsRequest{
 			SandboxId: sandboxID,
 			Name:      build.Name,
-		})); err != nil {
+		}))
+		if err != nil {
 			s.destroySandbox(ctx, agent, sandboxID)
 			s.failBuild(ctx, buildID, fmt.Sprintf("flatten rootfs failed: %v", err))
 			return
 		}
+		sizeBytes = flatResp.Msg.SizeBytes
 	}
 
 	// Insert into templates table as a global (platform) template.
@@ -297,7 +302,7 @@ func (s *BuildService) executeBuild(ctx context.Context, buildID string) {
 		Type:      templateType,
 		Vcpus:     pgtype.Int4{Int32: build.Vcpus, Valid: true},
 		MemoryMb:  pgtype.Int4{Int32: build.MemoryMb, Valid: true},
-		SizeBytes: 0, // Could query the host, but the template is created.
+		SizeBytes: sizeBytes,
 		TeamID:    platformTeamID,
 	}); err != nil {
 		log.Error("failed to insert template record", "error", err)
@@ -319,7 +324,8 @@ func (s *BuildService) executeBuild(ctx context.Context, buildID string) {
 }
 
 func (s *BuildService) waitForHealthcheck(ctx context.Context, agent buildAgentClient, sandboxID, cmd string) error {
-	deadline := time.After(healthcheckTimeout)
+	deadline := time.NewTimer(healthcheckTimeout)
+	defer deadline.Stop()
 	ticker := time.NewTicker(healthcheckInterval)
 	defer ticker.Stop()
 
@@ -327,7 +333,7 @@ func (s *BuildService) waitForHealthcheck(ctx context.Context, agent buildAgentC
 		select {
 		case <-ctx.Done():
 			return ctx.Err()
-		case <-deadline:
+		case <-deadline.C:
 			return fmt.Errorf("healthcheck timed out after %s", healthcheckTimeout)
 		case <-ticker.C:
 			execCtx, cancel := context.WithTimeout(ctx, 10*time.Second)
@@ -366,8 +372,11 @@ func (s *BuildService) updateLogs(ctx context.Context, buildID string, step int,
 	}
 }
 
-func (s *BuildService) failBuild(ctx context.Context, buildID, errMsg string) {
+func (s *BuildService) failBuild(_ context.Context, buildID, errMsg string) {
 	slog.Error("build failed", "build_id", buildID, "error", errMsg)
+	// Use a detached context so DB writes survive parent context cancellation (e.g. shutdown).
+	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
+	defer cancel()
 	if err := s.DB.UpdateBuildError(ctx, db.UpdateBuildErrorParams{
 		ID:    buildID,
 		Error: pgtype.Text{String: errMsg, Valid: true},
@@ -376,7 +385,10 @@ func (s *BuildService) failBuild(ctx context.Context, buildID, errMsg string) {
 	}
 }
 
-func (s *BuildService) destroySandbox(ctx context.Context, agent buildAgentClient, sandboxID string) {
+func (s *BuildService) destroySandbox(_ context.Context, agent buildAgentClient, sandboxID string) {
+	// Use a detached context so cleanup succeeds even during shutdown.
+	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
+	defer cancel()
 	if _, err := agent.DestroySandbox(ctx, connect.NewRequest(&pb.DestroySandboxRequest{
 		SandboxId: sandboxID,
 	})); err != nil {