dodjdnh opened a new issue, #1572: URL: https://github.com/apache/dubbo-admin/issues/1572
## Problem The official `release/kubernetes/` manifests already include ConfigMaps for six Dubbo metric and trace dashboards, a Grafana dashboard provider, a Jaeger Service, and support for dashboard URLs in Dubbo Admin. However, three provisioning connections are incomplete: 1. The six dashboard ConfigMaps are not mounted by the Grafana Deployment under the provider's scan path, `/grafana-dashboard-definitions/0`. The presence of dashboard JSON therefore does not mean the dashboards are automatically provisioned. 2. `grafana-dashboardDatasources.yaml` provisions only Prometheus. The Jaeger datasource required by the three trace dashboards is missing. Grafana should query Jaeger at `http://jaeger.monitoring.svc:16686`, not at the OTLP ingestion ports `4317/4318`. 3. Kubernetes Dubbo Admin starts with the ConfigMap-provided `/etc/dubbo-admin/dubbo-admin.yaml`, but that configuration lacks the six `console.metricDashboards` and `console.traceDashboards` URLs. Even if Grafana has the dashboards, Admin cannot generate their entry points from this Kubernetes configuration. ## Origin and history These are initial or later integration omissions, **not regressions where a previously correct implementation was removed**: | Gap | Introducing PR | Historical finding | | --- | --- | --- | | A: six dashboards are not mounted | [#1387](https://github.com/apache/dubbo-admin/pull/1387), merge commit `2eb4409b03420f687416763770c9b4e88ed2d0da` | `INITIAL_OMISSION`: that PR introduced the Grafana Deployment, provider, and six Dubbo dashboard ConfigMaps, but its first Deployment did not include their volumes and volume mounts. | | B: Jaeger datasource is missing | Also [#1387](https://github.com/apache/dubbo-admin/pull/1387) | `INITIAL_OMISSION`: that PR added the Jaeger Deployment/Service and trace dashboards that depend on Jaeger, while datasource provisioning contained only Prometheus from the outset. | | C: six Kubernetes Admin URLs are missing | [#1410](https://github.com/apache/dubbo-admin/pull/1410), merge commit `35cbd342b75467cc87f0f6763f72e8752d09fe47` | `LATER_PARTIAL_INTEGRATION`: the local example in #1387 already had the six URLs, and [#1397](https://github.com/apache/dubbo-admin/pull/1397) changed them to the current `/d/<UID>?kiosk=1&theme=light` form. When #1410 later established the actual Kubernetes Admin ConfigMap, mount, and `-c` loading path, it did not carry over those existing settings. | There was no actual Kubernetes Admin ConfigMap in #1387, so gap C should not be described as "#1387 removed URLs from the ConfigMap." #1410 did not change the Grafana monitoring manifests, so gaps A/B should not be attributed to #1410. ## Expected behavior - Mount all six existing Dubbo dashboard ConfigMaps into the existing Grafana provider scan path through the Deployment's volumes and volume mounts, so versioned manifests provision them automatically without manual UI import. - Keep Prometheus available and also provision a `type: jaeger`, `access: proxy` datasource pointing to the Jaeger query Service, so all three trace dashboards can resolve it. - Provide the Application, Instance, and Service metric and trace dashboard URLs in the Kubernetes Admin ConfigMap. All six UIDs should match the formal dashboard JSON, and the Grafana base URL should use the correct Kubernetes Service DNS name. ## Reproduction and scope At official baseline `bb906476e1aa923d957e6420560ca13aba17524f`, inspect the dashboard ConfigMap → Deployment mount → provider path, the Grafana datasource Secret, and the Admin ConfigMap → `-c` path to see gaps A/B/C. On the unfixed baseline with D002 applied, an external read-only Provisioning Oracle reported A1–A6 FAIL; B1 PASS, B2–B4 FAIL; C1–C6 FAIL; and `OVERALL: FAIL`. This issue concerns Kubernetes provisioning only. It does not require every metric or trace business query to pass, and does not include dashboard query changes, Instance Trace instance identity, the OTel Collector, Log, or cross-signal navigation. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
