aicoe / idh-manifests Goto Github PK

As Data Scientist/acting Software Engineer,

I would like to have an image that is based on common structure regarding Data Science project (e.g cookicutter) and some utils/functionalities that can be reusable (e.g. reading/storing file on Ceph, connect to Spark), either source code or template notebooks.

Having one git repository build with those requirements can be built automatically with the CI and we could all contribute to have not only a structure but common functionalities in order to avoid reinventing the wheel.

References:

• https://drivendata.github.io/cookiecutter-data-science/
• https://github.com/aicoe-aiops/project-template

Once Argo 2.12.0 is finally out (with RBAC and stuff), work together with Thoth to get the images updated (maybe move the quay repos from thoth-station to aicoe?) and update the deployment state from the current inconsistent state where 2.11.1 manifests are used with 2.11.0 images.

Reference: #55

/assign @tumido

Right now we dont have any authentication for superset except the admin user. It would be nice to have oauth support for superset so we can integrate it at the enterprise scale.
Currently, we are using LDAP but OAUTH will be a much more seamless experience across all other ODH components.

we would like request to use JupyterHub images from thoth-station quay so that users can have the latest version available.
Please include the following images, rather than building them during deploy of JupyterHub:

https://quay.io/repository/thoth-station/s2i-tensorflow-notebook
https://quay.io/repository/thoth-station/s2i-scipy-notebook
https://quay.io/repository/thoth-station/s2i-minimal-notebook

Once we have bumped up the available resources on the monitoring namespace we should increase the number of pods available for both prometheus and alertmanager

HA for Argo:

• https://argoproj.github.io/argo/tls/
• https://argoproj.github.io/argo/high-availability/

Resolve ERR_INCOMPLETE_CHUNKED_ENCODING issue:

http://argo-server-aiops-prod-argo.apps.ocp4.prod.psi.redhat.com/api/v1/workflow-events/aiops-prod-argo?listOptions.fieldSelector=metadata.name=anomaly-cron-wf-1612332600&listOptions.resourceVersion=0

sometimes results in: ERR_INCOMPLETE_CHUNKED_ENCODING