SEZ9 commented on issue #12630:
URL: https://github.com/apache/seatunnel/issues/12630#issuecomment-6030766796

   Thanks for the proposal — the direction (deterministic rules over fields the 
engine already publishes, a single read-only request, non-zero exit only for a 
terminal `FAILED` job) fits well with how we want Zeta operational tooling to 
behave. Before the implementation can be treated as ready, I'd like to see the 
three points raised above closed:
   
   1. **Missing-job handling.** If the server answers a job id it no longer 
retains with an HTTP 200 and a minimal body rather than a 404, a 404-only test 
won't exercise the real missing-job path and the command could print `UNKNOWN` 
and exit 0. Please confirm what the server actually returns in that case, treat 
the minimal success payload as "no retained job" with a non-zero result, and 
add a loopback test for that exact response.
   
   2. **Cost of `/job-info`.** As noted above, that endpoint may restore the 
logical DAG and, on a cache miss, fall through to a slower, nondeterministic 
path. The diagnose command only needs status, diagnostics timestamps, restore 
counters, source/sink counts and the error message. Please either state 
explicitly in the proposal that v1 deliberately pays for the full `job-info` 
path, or sketch a bounded lightweight diagnostics response, before describing 
the command as harmless on a busy cluster.
   
   3. **REST client and error redaction.** Keep the REST-client extraction in 
its own change rather than duplicating it here, and make sure the 
error-response body passes through credential redaction before being displayed, 
since it is opaque server text. Please add a regression for a non-404 error 
response whose body contains a password, asserting the secret is absent while 
the error code stays visible.
   
   Once those are addressed, the rule and exit-code tests give us a good base 
to review the actual diagnosis behaviour. Could you update the proposal text 
here with your decision on point 2 and link the follow-up changes for points 1 
and 3?
   
   <!-- streview-comment:1580 -->


-- 
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