* add more option for controller
1. add ConcurrentReconciles for setting the concurrent reconcile number of the controller
2. add DependCheckWait for setting the time to wait for ApplicationConfiguration's dependent-resource ready
* fix test
* add controller reference
solve failed test
add compatibility test for old crd
add app ns for cli
modify compatibility test
solve compatibility problem
add testing for GetDefinition func with cluster scope CRD
generate code for compatibility-test
move testdata generate to makefile
optimize ci pipeline for compatibility-test
* adjust CI scripts
* fix go ci order
* try fix
* use appfile to update trait
* fix vela status
* use assert.ObjectAreEqual to check diff for map case
* remove application controller own AC and Components
* tune place
* fix lint
* update
* remove error check of vela init test
Co-authored-by: Ryan Zhang <yangzhangrice@hotmail.com>
* Generate restful api based on Swagger
As OpenAPI has to be quickly developped to support front-end developement, so
did the restufl api docs. Based on swagger, generate the docs automatically
* fix conflicts and add application list annotation
Automatically generate reference docs for all workloads and traits.
1) moved design part from hack/references/configurations/rollout.yaml to design/vela-core/rollout.md
2) added usages for all workloads/traits, except route (#672)
3) renamed some reference docs
Attention: atuo-generation of specifications for all parameters has NOT been implemented, they are generated
based on files under `hack/references/configurations`
Co-authored-by: Jianbo Sun <wonderflow.sun@gmail.com>
currently the charts has following issues:
- cert-manager has manifests in multiple places. Should be combined into one.
Prefer pure yaml since its helm chart requires `--set installCRDs=true` flag and uninstalling CRD via helm is inconvient.
- The bootstrap of vela-core will setup metrics controller. This requires Prometheus Operator CRD. Move related CRD to vela-core chart.
- move prometheus to vela-system namespace. Remove monitoring namespace.
- remove grafana. It's already a dependency of prometheus chart.
- change default image pull policy to IfNotPresent. Should not be Always.
Signed-off-by: Hongchao Deng <hongchaodeng1@gmail.com>
yangsoon
Weiping Cai
wangyike
zzxwill
roy wang
Jianbo SunandHongchao Deng
Ryan Zhang
majian
guoxudong
