You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is an RFE for how we want to design our separation of public, private and authenticated API RBAC model after we migrate fully from CherryPy to Flask API endpoints.
Provide basic public API access to tenant automation about their machines
Provide web-based API access to endpoints that help build our landing pages, requests, wiki and status pages to illustrate more up-to-date and granular systems, assignment and availability information
Protect API endpoints that should be private or local (re-architect them to use localsocket only for example)
Extend capabilities in other areas
Design
We want to ensure any local API calls are done against gunicorn listening on localhost as well, skipping the nginx reverse proxy tier while ensuring all public API calls utilize nginx reverse proxy tier and can reap all the benefits that gives us.
As a first pass, here might be some of the API classification we'd want to do:
(brainstorm) We might allow a subset of API calls for tenants to interact with their hosts after they receive them, this might require the implementation of [RFE] LDAP integration #191 first.
The text was updated successfully, but these errors were encountered:
Resolving this as we mostly accomplished everything with release of 2.0.0.
I think that we'll need to further granularize concept of users as generic cloud environments but this can be separate RFE's as needed to support: #487
Summary
This is an RFE for how we want to design our separation of public, private and authenticated API RBAC model after we migrate fully from CherryPy to Flask API endpoints.
Depends on:
Related RFE
Details
We'll want to separate our API endpoints (prominent endpoints documented here: https://github.com/redhat-performance/quads/blob/master/docs/quads-api.md ) into a more role-based or purpose-driven separation design for the following reasons below.
Why?
Design
We want to ensure any local API calls are done against gunicorn listening on localhost as well, skipping the nginx reverse proxy tier while ensuring all public API calls utilize nginx reverse proxy tier and can reap all the benefits that gives us.
As a first pass, here might be some of the API classification we'd want to do:
Open / Public API Calls
GET
API endpoints, mostly documented here: https://github.com/redhat-performance/quads/blob/master/docs/quads-api.md#api-get-operationsProtected / Local API Calls
POST
API calls, the main ones documented here https://github.com/redhat-performance/quads/blob/master/docs/quads-api.md#api-post-operationsAuthenticated / Protected Public API Calls
The text was updated successfully, but these errors were encountered: