X-LightYear opened a new issue, #7338:
URL: https://github.com/apache/rocketmq-dashboard/issues/7338
## Problem
The paginated consumer-group endpoint is not provider-native for Aliyun and
Tencent instances.
`MetadataService.listConsumerGroupsPage` delegates to
`InstanceProvider.listConsumerGroupsPage`,
but neither cloud provider overrides the default implementation. The default
first loads and
enriches the complete consumer-group inventory, then slices one page in
memory.
This makes a request such as `GET /api/groups/page?page=1&pageSize=20`
perform work proportional
to every matching group. On Tencent, the full listing also enriches each
group through a detail
call, so the page request can fan out to one additional remote call per
group before returning
the first page.
## Expected behavior
Aliyun and Tencent paginated consumer-group requests should use the provider
OpenAPI page
parameters and the service-reported total when available. Only the requested
page should be
converted and enriched. Existing search behavior and the fallback path for
responses without a
usable total should remain unchanged.
## Actual behavior
Aliyun and Tencent inherit the generic in-memory pagination fallback. The
full inventory is fetched
and enriched before the requested page and total are produced.
## Reproduction
1. Configure an Aliyun or Tencent RocketMQ instance with more consumer
groups than one page.
2. Open the consumer-group page or call `GET /api/groups/page` with a small
`pageSize`.
3. Observe that the provider first enumerates the complete inventory rather
than requesting the
requested page from the cloud API.
The deterministic regression tests added for this change reproduce the
contract with a response
reporting `totalCount=201` for page 3. Before the fix, Aliyun returns the
fallback total `1` and
an empty page; Tencent throws `BusinessException: Tencent Cloud returned an
incomplete consumer
group page` from the full-list path.
## Scope
Implement native consumer-group page handling for Aliyun and Tencent while
preserving the generic
fallback for unsupported or incomplete provider responses. The existing
Apache/database-backed
pagination work is outside this change.
## Validation
- Aliyun and Tencent provider regression tests assert page parameters,
returned items, and totals.
- Pre-fix baseline: both regression tests fail for the reasons above.
- Post-fix provider tests: 132 passed.
- Provider plus MetadataService tests: passed.
- Checkstyle: passed with 0 violations.
- Server package compilation completed; the environment's existing `python3`
binary-license gate
is unavailable on Windows (exit code 9009).
--
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]
