simple-task-manager 1.6.1

Simple task manager to handle execution of tasks in AWS docker or any system. It use a handler that can be loaded dynamically.

Simple Task Manager (STMGR)

This is a simple task manager to start processes on AWS/Docker or other platforms. The app includes classes to manage tasks on AWS, but you can implement additional handlers dynamically.

You only need to create a class to handle the task that you want to start and register it in a database (DDBB). Below, I'll compare STMGR with Celery (since it's widely used) to explain the key differences and help you make an informed choice. Another good choice is Dramatiq. For a simple comparison between STMGR and Celery check

Requirements

• Python 3.10 or newer
• PostgreSQL
• psycopg[binary]

Installation

This project can be installed using pip:

pip install simple-task-manager

Or it can be installed directly from git: pip install git+https://github.com/Fran-4c4/staskmgr

Upgrade from 1.5.x

• Existing client code using TaskDB can remain synchronous.
• The service runtime now uses an async SQLAlchemy layer internally.
• Backward compatibility is preserved through compatibility modes:
  • LEGACY: old behavior only
  • AUTO: use safe mode when the upgraded schema is available
  • SAFE: require the upgraded schema and enable lease-based recovery
• To upgrade an existing installation, apply config/ddbb_upgrade_1_6.sql before enabling SAFE mode.

See Upgrade guide.

Task log correlation

Version 1.6.1 adds best-effort task correlation metadata for external containers. TASK_ID remains the minimum identity needed to execute a concrete task; optional fields such as PARENT_REQUEST_ID, PROCESS_CHAIN_ID, TASK_TYPE, TASK_MANAGER, SERVICE_NAME and DEPLOYMENT_ENVIRONMENT are propagated when available.

Docker launches receive environment variables and labels. ECS launches receive container environment overrides and task tags. Missing optional metadata does not break legacy task definitions.

Usage and requirements

• First you need to configure the minimum parameters in order to run tasks. See Configuration
• Second you need a database to store configuration and task management. See table creation in folder config/ddbb_script.sql or Configuration scripts. Actually only PostgreSQL is supported.

More info in github GitHub.

---

Adding handlers

In order to manage other types you need to create a class and an entry in DDBB or in your appconfig.json in the section task_handlers. When the task is retrieved from DDBB it will look the handler. Below is an example of the Test task handler.

"task_handlers": {
    "TEST_MGR": {
      "config": {
        "task_handler": {
          "name": "TestTaskHandler",
          "path": "task_handlers",
          "class": "TestTaskHandler",
          "module": "test_task_handler",
          "launchType": "INTERNAL",
		      "task_next_status":"FINISHED"
        }
      }
    }
  }

Test in local

Install using pip in your project using the next command and changing x.x.x version.

pip install "path_to_dist/dist/Simple_Task_Manager-x.x.x-py3-none-any.whl" 

Sphinx documentation

• There is a minimal documentation generated from source. See Documentation

Running as a Docker service

STMGR is normally deployed as a service inside Docker.

1. Build an image based on Python 3.10 or newer.
2. Install the package version you want to run.
3. Mount the configuration and provide database credentials.
4. Apply the SQL upgrade before enabling SAFE mode.

See Docker notes.

Docker handler modes

DockerTaskHandler now supports two modes:

• DETACHED: default mode, intended for heavy containers. Launch and return immediately.
• BLOCKING: wait for the container to finish. Intended for lightweight containers.
• In DETACHED, STMGR can later reconcile the container state using external_ref when the upgraded schema is enabled.

See Docker handler guide.

License

licensed under Apache License 2.0