Nano-Run

Lightweight async request runner. GitOps friendly.

A simplified version of trusted-cgi designed for async processing extreme amount of requests.

Main goals:

• Should have semi-constant resource consumption regardless of:
  • number of requests,
  • size of requests,
  • kind of requests;
• Should be ready to run without configuration;
• Should be ready for deploying in clouds;
• Should support extending for another providers;
• Can be used as library and as a complete solution;
• Performance (throughput/latency) has less priority than resource usage.

Please note that the project is being developed in free time, as a non-profitable hobby project. All codes, bugs, opinions, and other related subjects should not be considered as the official position, official project, or company-backed project to any of the companies for/with which I worked before or/and at present.

TODO:

• On attempt failed job
• On failed job
• On success job
• Per endpoint swagger
• Pickup file as an artifact

Documentation

• Installation
• Quick start
• Architecture overview
• API
• API Authorization
• UI
• UI Authorization
• Unit configuration
• Docker
• Cron-like scheduler

Stability

(After 1.0.0)

We are trying to follow semver:

• Patch releases provides fixes or light improvements without migration
• Minor releases provides new functionality and provides automatic migration if needed
• Major releases provides serious platform changes and may require manual or automatic migration

Within one major release, it guarantees forward compatibility: new versions can use data from previous versions, but not vice-versa.

Reproducible binaries

The project tries to follow best practices providing reproducible binaries: it means, that you can verify that complied binaries will be exactly the same (byte to byte) as if you will compile it by yourself by following our public instructions.

License

The project (source code and provided official binaries) are licensed under Apache-2.0 (see License file, or in plain English) and suitable for personal, commercial, government, and others usage without restrictions as long as it used with abiding license agreement.

Do not forget to bring your changes back to the project. I will be happy to assist you with PR.

The project uses external libraries that may be distributed under other licenses.

Installation

Debian/Ubuntu

(recommended)

Tested on 20.04 and 18.04, but should good for any x64 version.

Add the repository (only once)

sudo apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv-keys 379CE192D401AB61
echo "deb https://dl.bintray.com/reddec/debian all main" | sudo tee /etc/apt/sources.list.d/reddec.list

Install or update nano-run

sudo apt update
sudo apt install nano-run

Automatically creates service nano-run.service.

Binary

Download and unpack desired version in releases.

Docker

docker pull reddec/nano-run

From source

Requires go 1.14+

go get -v github.com/reddec/nano-run/cmd/...

Ansible for debian servers

- name: Add an apt key by id from a keyserver
  become: yes
  apt_key:
    keyserver: keyserver.ubuntu.com
    id: 379CE192D401AB61
- name: Add repository
  become: yes
  apt_repository:
    repo: deb https://dl.bintray.com/reddec/debian all main
    state: present
    filename: reddec
- name: Install nano-run
  become: yes
  apt:
    name: nano-run
    update_cache: yes
    state: latest

Quick start

(optional) initialize configuration

nano-run server init

it will create required directories and files in a current working directory.

define a unit file

Create minimal unit file (date.yaml) that will return current date (by unix command date) and put it in directory run/conf.d/

run/conf.d/date.yaml

command: date

start nano-run

nano-run server run

now you can open ui over http://localhost:8989 or do API call: curl -X POST http://localhost:8989/api/date/