JAC Cloud Orchestrator

[AshishMahendra]

Description

JAC Cloud Orchestrator is designed to dynamically import any Python module, deploy it as a Kubernetes Pod, and expose it as an independent gRPC service. This enables any Python module to be a microservice, providing flexibility and scalability in a cloud environment.

[AshishMahendra]

Hi, @marsninja, can you please share your thoughts on the approach, i have added a readme for reference

[marsninja]

Good start @AshishMahendra, Some of this implementation looks relevant to the final approach.

For the first version of jac-cloud-orc (which by the way im thinking if jac-splice-orc might be better) would be the following:

• A plugin that overrides jac_import to add capability that will automatically turn particular named modules into remote containerized kubernetes pod services that the importing modules would send all interactions to
• Initially these modules can be specified in a config file or through the settings.py interface
• This plugin would use the kubernetes python client library to automatically detect whether these running jac application is in a kubernetes environment and has the right permissions to create pods in that environment (this was prototyped in jaseci 1 and is do-able)
• The plugin would automatically generate the relevant manifests through the k8s protocol to start the pod and initially stall until the pod was ready, then the rest of the module can then execute.

Basically it would be something like that, i'd want as much automatic control being done by splice-orc as possible.

[marsninja]

Hi @AshishMahendra I see you've assigned this to me. What feedback do you want? Is it ready for review and potential merge?