/docker-flow-cron

Primary LanguageGo

Docker Flow Cron

This project is in the design phase.

High Level Design

The project should be able to:

  • schedule jobs using cron syntax
  • runs jobs detached from the cron so that they are unaffected in case of a failure
  • run jobs as Swarm services
  • schedule jobs through HTTP requests
  • use constraint, reserve-cpu, reserve-memory, and label arguments
  • rerun failed jobs
  • retrieve the list of scheduled jobs
  • retrieve job details
  • retrieve job executions
  • retrieve job execution logs (docker service logs)
  • remove scheduled jobs
  • update scheduled jobs
  • be fault tolerant
  • after cron rescheduling, it does not start jobs that are already running
  • send notifications when a job fails
  • provide dashboard for job monitoring (explore the option to extend Portainer or Grafana)
  • time out jobs
  • provide UI for administration (explore the option to extend Portainer)
  • host pre-made jobs (e.g. docker prune, Mongo backup, etc)
  • prune old executions (services)
  • create a CLI
  • provide commands that will allow users to filter services with Docker client

Tasks

  • develop PoC
  • document (use cron.AddJob as the base for the rules)
  • release
  • reduce testing time or split into groups (e.g. unit tests, integration tests)

Assumptions

  • Docker experimental features are enabled while docker service logs is in it
  • System has /bin/sh
  • Binary is running on a manager node