unbridled-41 opened a new pull request, #4748:
URL: https://github.com/apache/rocketmq-dashboard/pull/4748
Fixes #4747.
## Problem
One enabled alert rule stored without a metric makes the alert
reconciliation pass throw, so every active alert of that instance and domain
stays active forever — no `RESOLVED` state, no `RESOLVED` event, no recovery
notification.
- `ops/alert/NativeAlertProcessor.java:113-116` filtered the domain's rules
with
`scope.metricKeys().contains(StringUtils.trimWhitespace(rule.getMetric()))`.
- `MetricCollectionScope` holds its keys as `Set.copyOf(...)`
(`cluster/metrics/MetricCollectionScope.java:31`); every `Set.of`/`Set.copyOf`
implementation rejects a null lookup (`Set12.contains` dereferences the
argument, `SetN.contains` calls `Objects.requireNonNull`), and
`StringUtils.trimWhitespace(null)` is null.
- A metric-less rule is storable: `AlertRuleRequestDTO.metric`
(`ops/alert/AlertRuleRequestDTO.java:34`) has no `@NotBlank`, the column is
nullable (`db/schema.sql:274`), and the JSON import path tolerates it
explicitly (`AlertRuleTransferService.java:33-46`,
`NativeAlertMetricCatalogService.java:70-73`, reachable from
`AlertRuleController.java:85-86`).
- The failure is swallowed: `CollectorScheduler.java:282-284` / `:298-300`
catch `RuntimeException` and log a warning, so the reconcile pass for the whole
scope is skipped.
## Root cause
A null-unsafe membership test over an immutable set, in a filter that runs
before any metric-less rule is discarded — so the rule is not skipped, it
aborts the pass.
## Fix
`NativeAlertProcessor` normalises the rule's metric to `""` when it is
absent, so the membership test answers `false` and the rule is filtered out
while the rest of the scope is still reconciled (`normalizedMetric`, 13 lines
including the comment). A rule without a metric can never match a collected
sample, so it has no active state that the pass could resolve; the semantics of
every other rule in the scope are unchanged.
## Tests
Red before the fix:
```
[ERROR]
NativeAlertProcessorTest.reconcilesRemainingRulesWhenAnotherRuleHasNoMetricTest
-- Time elapsed: 0.016 s <<< FAILURE!
"java.lang.NullPointerException: Cannot invoke "Object.equals(Object)"
because "o" is null
[ERROR] Tests run: 24, Failures: 1, Errors: 0, Skipped: 0
```
Commands and results after the fix, run from `server/`:
- `mvn -B -ntp -o test -Dtest=NativeAlertProcessorTest` → **24 passed**,
`BUILD SUCCESS`.
- `mvn -B -ntp -o test
-Dtest='org.apache.rocketmq.studio.ops.alert.**,CollectorSchedulerTest'` →
**318 tests, 0 failures, 5 errors**; all 5 errors are the sandbox's missing
MySQL (`CannotGetJdbcConnectionException: Communications link failure`, from
`NativeAlertEvaluationTransactionTest`,
`NotificationOutboxMapperIntegrationTest`,
`RmqAlertStateMapperIntegrationTest`), the same baseline failure this
environment produces on the clean trunk.
## Duplicate check
Searched open PRs (205) and issues (214) for `reconcile`, `null metric`,
`without a metric`, `resolve`, `alert state`, plus the commit index: the
adjacent open items are #4571 (the AI rule-list tool tolerating a metric-less
rule) and #4711 (rejecting a metric-less rule *test run*) — same input class,
two other surfaces, neither touching
`NativeAlertProcessor.reconcileMissingActiveStates`. No open item covers this
crash site. #4647/#4648 touch a different unguarded parse in the same scheduler.
## Base branch
Targets the development trunk `rocketmq-studio`. GitHub interprets closing
keywords only for pull requests that target the repository default branch, and
that is still `master`, so merging will not close #4747 automatically — it
needs to be closed by hand.
## Risk
Low and narrow: the filter now skips a rule that previously made the pass
abort before doing anything. No behaviour is removed for rules that carry a
metric, and the change cannot turn a resolvable alert into an unresolvable one.
--
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]
