renelick 0.25.2rc2

Data-Driven Task Orchestration

Renelick

This project is to provide low-level distributed task orchestration. It was forked from the KernelCI API code base to evolve into a more generic implementation than the particular use-case of automated Linux kernel builds and runtime tests on devboards.

Renelick's key features are:

• Data tree and pub/sub events

  All the data is stored in a tree of Nodes objects which can hold any application-specific attributes.
  Whenever some data gets added, a pub/sub event is generated which can be used as a trigger for automation.

• Orchestration and message queues

  Independently from the data, tasks can be scheduled via a distributed orchestration mechanism.
  Orchestrators can schedule tasks following particular triggers using message queues for load balancing.

• User management

  Users can register with email verification, authenticate with OAuth2, get temporary JWT tokens and persistent API keys.

• Runtime abstraction

  Runtime environments can be deployed anywhere, from standard Kubernetes to private instances with specific hardware.

• Modular web dashboard

  The provided dashboard covers the common parts and can be extended with React elements for specific applications.

While the design principles are now well established, it is still under active development. The approach is to iterate over the whole stack, alternating between design decisions and implementation. Each iteration or "sprint" has a version number and an associated milestone.

Please see the documentation for full details, and the quick-start guide for setting up a local development instance.