iloveeyjafjalla opened a new issue, #6298: URL: https://github.com/apache/rocketmq-dashboard/issues/6298
### Before Creating the Bug Report - [x] Searched all-state Issues/PRs and inspected related current full head differences and all public comments. - [x] This is a RocketMQ Studio frontend issue in this repository. - [x] Reproduced on the exact revision below. ### Studio Version `rocketmq-studio@5e4c39b053c35a5598cc79ff331b8a54e649d7b1`. ### Runtime Environment / Build Toolchain Windows 11, Node 24.15.0, npm 11.12.1. Automated reproduction uses the actual AlertsPage rule editor and controlled ordinary service responses. No deployed incident or live metric collection is claimed. ### Connected RocketMQ Cluster No broker/database contacted. The affected normal frontend path validates the current editor and calls the rule Test Run service. ### Describe the Bug Inside one open rule editor, changing evaluation inputs does not clear a completed dry-run table or invalidate its pending request. `handleTest` stores and announces the response without checking whether the tested criteria remain current. The table has no tested-threshold snapshot or outdated-result marker. ### Steps to Reproduce 1. Edit an ordinary BUSINESS `consumer.lag.total` rule for a consumer group, operator `>`, threshold 1000 messages. 2. Click Test Run. Let a valid available sample with value 1200 and `conditionMet: true` return and display Matched. 3. Change the actual threshold input to 2000 in the same editor. 4. The old table still displays value 1200 as Matched, with no indication it belongs to the previous threshold. The second deterministic regression changes the threshold while the first test is pending, then releases that old response: its table is restored against the new editor criterion. Neither case requires switching rule, domain or route, malformed input, or a fabricated broker incident. ### What Did You Expect to See? Results calculated for changed criteria should be cleared or explicitly marked outdated. A pending obsolete response should not restore them or interfere with a newer test's result/loading/feedback. Metadata-only edits need not invalidate unchanged evaluation criteria. ### What Did You See Instead? Old samples and their verdict persist or return after an input change. The real-page regressions fail on the visible old table; no threshold edit triggers a new test request automatically. ### Additional Context #4583 / merged #4584 clears completed samples when opening another rule. #6185 / current #6190 isolates domain/editor sessions. Their complete source/test differences and public comments do not protect criterion edits within the same still-open editor; this report addresses that distinct reproduced trigger. The proposed fix invalidates by evaluation fields and guards obsolete continuations, without calculating alert conditions in the frontend or changing backend/API semantics. Actual validation and limits are recorded in the PR. AI assistance: OpenAI Codex investigated, implemented and tested the change with independent automated review; no human or maintainer approval is claimed. ### Are You Willing to Submit a Pull Request? - [x] Yes. ### Implementation PR https://github.com/apache/rocketmq-dashboard/pull/6297 -- 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]
