docs(test-plans): replace removed /providers API with /agents and /catalog-items - #38
Conversation
Code Review by Qodo🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0)
Great, no issues found!Qodo reviewed your code and found no material issues that require reviewTip of the day💡 Did you know, you can copy the agent prompt from any finding and feed it to your IDE agent |
PR Summary by QodoMigrate authentication test plan from providers to agents
AI Description
Diagram
High-Level Assessment
Files changed (1)
|
2ea37fb to
59d0c14
Compare
…talog-items control-plane#51 removed the /providers endpoint entirely (replaced by agent-based architecture over NATS). Update the FLPATH-3254 auth test plan: - TC-08: rewrite from "Provider CRUD" to "Agent CRUD" (POST/GET /agents) - TC-02, TC-34: replace provider create/delete with agent registration - All other TCs: /providers → /catalog-items as generic protected endpoint - TC-37/38: clarify step titles for SP-as-agent context - Add migration note documenting the API change Signed-off-by: Vladislav Kolodny <vkolodny@redhat.com> Co-authored-by: Cursor <cursoragent@cursor.com> Signed-off-by: Vladislav Kolodny <vkolodny@redhat.com> Co-authored-by: Cursor <cursoragent@cursor.com>
59d0c14 to
9c3e00d
Compare
|
Code review by qodo was updated up to the latest commit 9c3e00d |
| #### Cleanup | ||
|
|
||
| None - provider deleted in step 5. | ||
| No DELETE endpoint for agents. Test agent remains in DB; subsystem tests use a fresh Postgres volume per run. |
There was a problem hiding this comment.
However, this is a MANUAL test case (not just subsystem), so testers running this locally will accumulate test agents. The cleanup should provide a SQL DELETE statement or explain that running make compose-down -v removes the agent. The current text only addresses automated subsystem test runs. Can you double check?
| **Expected:** Token obtained. | ||
|
|
||
| **Step 2: Create a provider** | ||
| **Step 2: Register an agent** |
There was a problem hiding this comment.
This step egisters an agent named 'helm-smoke-agent' but doesn't verify whether an agent with that name already exists from a previous test run. Since there's no DELETE endpoint and agents persist, re-running this test will trigger the idempotent update path (HTTP 200) instead of creation (HTTP 201), causing the expected result to be wrong. Can you please double check?
| curl -s -H "Authorization: Bearer $TOKEN" http://localhost:8080/api/v1alpha1/agents | jq '.agents[] | {name, health_status}' | ||
| ``` | ||
|
|
||
| **Expected:** Registered SPs report ready/healthy via `health_status` (or documented known-unhealthy exceptions such as ACM SP GVK scheme issues tracked separately). |
There was a problem hiding this comment.
the description should be updated to reflect that we're polling agent health status, not provider health directly no?
- TC-08 cleanup: add SQL DELETE for manual testers + compose-down note - TC-08/TC-34: expected accepts 201 or 200 (idempotent re-registration) - TC-38: fix expected text to say "agents" not "SPs" Signed-off-by: Vladislav Kolodny <vkolodny@redhat.com> Co-authored-by: Cursor <cursoragent@cursor.com>
|
TC-02 Step 2 needs a quick update to match TC-08 and TC-34. Expected result currently only shows HTTP 201, but agent registration is idempotent by name — re-running returns HTTP 200. TC-08 and TC-34 already document this from earlier review feedback. Manual testers re-running TC-02 will see 200 and think it failed. |
Summary
control-plane#51 removed
the
/providersendpoint entirely (replaced by agent-based architecture over NATS).POST/GET /agents)/providers→/catalog-itemsas generic protected endpointNo functional test coverage lost — agent DELETE doesn't exist by design
(FK constraints preserve resource→agent history).
Test plan
/catalog-itemsand/agentsreturn 200 on live stack