GoogleContainerTools/skaffold · error

watching files for artifact %q: %w

Error message

watching files for artifact %q: %w

What it means

Dev() registers build dependencies for each artifact with the file monitor via r.monitor.Register. If computing or registering the watch set for an artifact fails, Skaffold emits DEVINIT_REGISTER_BUILD_DEPS and aborts with 'watching files for artifact <image>'. The failure comes from the artifact's BuildDependencies() (e.g. enumerating source files) or the monitor's registration.

Source

Thrown at pkg/skaffold/runner/dev.go:406

				func() ([]string, error) {
					return r.sourceDependencies.TransitiveArtifactDependencies(ctx, artifact)
				},
				func(e filemon.Events) {
					s, err := sync.NewItem(ctx, artifact, e, r.Builds, r.runCtx, len(g[artifact.ImageName]))
					switch {
					case err != nil:
						log.Entry(ctx).Warnf("error adding dirty artifact to changeset: %s", err.Error())
					case s != nil:
						r.changeSet.AddResync(s)
					default:
						r.changeSet.AddRebuild(artifact)
					}
				},
			); err != nil {
				event.DevLoopFailedWithErrorCode(r.devIteration, proto.StatusCode_DEVINIT_REGISTER_BUILD_DEPS, err)
				eventV2.TaskFailed(constants.DevLoop, err)
				endTrace()
				return fmt.Errorf("watching files for artifact %q: %w", artifact.ImageName, err)
			}
		}
	}

	// Watch test configuration
	for i := range artifacts {
		artifact := artifacts[i]
		if err := r.monitor.Register(
			func() ([]string, error) { return r.tester.TestDependencies(ctx, artifact) },
			func(filemon.Events) { r.changeSet.AddRetest(artifact) },
		); err != nil {
			event.DevLoopFailedWithErrorCode(r.devIteration, proto.StatusCode_DEVINIT_REGISTER_TEST_DEPS, err)
			eventV2.TaskFailed(constants.DevLoop, err)
			endTrace()
			return fmt.Errorf("watching test files: %w", err)
		}
	}

View on GitHub (pinned to a1189de023)

Solutions

1. Fix the artifact path/globs in skaffold.yaml so the build context resolves
2. Check the wrapped error for which file or directory is unreadable and restore permissions
3. Raise inotify watch limits: `sysctl fs.inotify.max_user_watches=524288`
4. Re-run `skaffold dev`

Example fix

// before (skaffold.yaml)
artifacts:
  - image: app
    context: ./src/app   # directory removed
// after
artifacts:
  - image: app
    context: .

// verify every artifact build context resolves before dev
for _, a := range cfg.Build.Artifacts {
    if _, err := os.Stat(a.Workspace); err != nil {
        return fmt.Errorf("artifact %q workspace missing: %w", a.ImageName, err)
    }
}

When it happens

Trigger: `r.monitor.Register(artifact.Dependencies, ...)` (inside RegisterBuildDependencies) returns an error for a specific artifact.ImageName — typically when resolving the artifact's build context file list fails (unreadable directories, bad sync/workspace globs) or the watcher cannot be created.

Common situations: Build context directory deleted or unreadable (permissions); Dockerfile referenced by the artifact missing; file watcher limit (inotify watches) exhausted; a workspace path in skaffold.yaml points outside the repo.

Related errors

• watching test files: %w
• watching skaffold configuration %q: %w
• unable to start trigger: %w
• didn't sync any files
• executing build: %w

AI-assisted analysis of GoogleContainerTools/skaffold@a1189de023 (2026-09-05). Data as JSON: /api/errors/555f7437bf104e4d. Report an issue: GitHub.