Table of Contents

Executable apps

An executable app gives a host-local command a stable place in the resource graph. It is useful for workers, tools, emulators, scripts, and services that do not need a language-specific provider.

Note

Executable apps are preview functionality for trusted hosts. Paths, arguments, and working directories are privileged host inputs and should not be offered to ordinary users of a shared CloudShell environment.

What you get

  • Command lifecycle. Start, stop, and restart the process from Resource Manager.
  • Stable resource identity. Keep endpoints, dependencies, logs, and activity attached to the command across restarts.
  • Process diagnostics. Capture standard output and error through the resource Logs view.
  • Flexible integration. Add environment variables, volume mounts, endpoint declarations, and references without writing a custom provider.

Minimal resource template

resources:
  - type: application.executable
    name: background-worker
    path: ./tools/worker
    command:
      arguments: --watch
      workingDirectory: ./tools

Paths are resolved on the CloudShell host. Use a plain command name when the executable should be resolved from the host's PATH.

When to choose it

Choose an executable app for a trusted local command that does not benefit from project-aware JavaScript, Java, Go, Python, or .NET behavior. Choose a container app when the workload needs a packaged boundary, replicas, or placement on a shared host.

Edit this page