SEZ9 commented on PR #11613:
URL: https://github.com/apache/seatunnel/pull/11613#issuecomment-5707788568

   Thanks for going through each point against `a947704b`.
   
   **F3 (auth on the two new endpoints)** — Thanks for the explanation of how 
both servlets sit on the same `ServletContextHandler` with `BasicAuthFilter` at 
the context level and mTLS at the connector layer. That resolves the concern. A 
short note in the docs for `/update-local-member-tags` and 
`/http-service/status` stating they use the same basic-auth/mTLS configuration 
as the other endpoints would be a nice addition.
   
   **F2 (RestApiIT/E2E coverage)** — Agreed it's non-blocking given 
`UpdateTagsServiceTest` and `HttpServiceStatusServletTest`. Ideally we'd still 
add one `RestApiIT` case hitting both new endpoints over real HTTP and one 
asserting the legacy flat map contract for `/update-tags`. If you'd prefer a 
follow-up PR, please note that in the description so we can track it.
   
   **F4 (local-only UUID validation behind a LB)** — Understood that 
`validateTargetMember()` is intentionally scoped to the node serving the 
request. The remaining ask is docs only: an explicit callout that the Web UI 
must reach the target member directly, and that a load balancer fronting 
multiple masters will cause the tag editor to reject updates for other members.
   
   **F6 (Restore Latest State on a RUNNING source job / cleaned-up 
checkpoint)** — Thanks for confirming this is open. Agreed it's not a 
regression from this PR and not a merge blocker. Before closing it I'd like one 
of: a server-side guard (reject while the source job is still RUNNING, and 
return a clear error when the referenced checkpoint state no longer exists), or 
an explicit "Limitations" note in the Web UI docs. A guard is preferred; a 
documented limitation is acceptable for this PR.
   
   **F1 / F5 / F7 / F8 (docs)** — Your last comment appears to be cut off at 
point 5, so I can't tell what was agreed. Could you confirm whether you plan to 
address in this PR: the misleading key name in the `/update-tags` example (F1); 
a dedicated `/update-local-member-tags` section with request, response and 
error examples (F5); per-field descriptions and a consistent example port for 
`/http-service/status` (F7); and updated screenshots for the new Action 
controls, Submit Job panel, Checkpoints tab and Operations page (F8)?
   
   Once the docs updates and either the F6 guard or the documented limitation 
are pushed, I'm happy to take another look.
   
   <!-- streview-comment:1114 -->


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to