use input files/directories -> requires a certain file naming strategy
...yaml
need to rework reconcilers and libraries -> to use the loaded input
use database -> strategy for upstream right now if we change this -> we take input and load this first + run reconciler -> in a hierarchical upstream this means we need to load all upstreams hierachically first and load the deepest one, etc etc etc
load as reconcilers/libraries files from the dev environment
TODO:
- load from dev environment (reconcilers and libraries)
- input loading -> there is no chnage -> only replace operation -> w/o metadata ->
- rework upstream loading with the hierarchical approach
directories:
- crds: remain seperate -> to be able to load apis, but apis should not dynamically load
- libs and reconcilers:
- they need to be preprocessed -> and put as input ()
- db: computed data
step1: choreoctl dev parse -> unifies reconciler/libraries in input directory step2: choreoctl server start or choreoctl server start + choreoctl server apply <ctx (url + ref + etc)> step3: chorectl run once or run start/stop
hierarchical loading of upstreams