📜 Description
DeleteSliceQoSConfig records two Prometheus counter metrics with incorrect event label values. At line 179, the metric is recorded with string(events.EventSliceConfigDeletionFailed) instead of string(events.EventSliceQoSConfigDeletionFailed). At line 190, it is recorded with string(events.EventSliceConfigDeleted) instead of string(events.EventSliceQoSConfigDeleted). The Kubernetes Event recorder calls on the same lines correctly use the QoSConfig-specific constants. Only the Prometheus event label is wrong, causing QoSConfig deletion metrics to be attributed to SliceConfig events in all dashboards and alerts.
👟 Reproduction steps
👟 Reproduction steps
- Read
service/slice_qos_config_service.go:174 — util.RecordEvent uses events.EventSliceQoSConfigDeletionFailed (correct).
- Read line 179 —
mf.RecordCounterMetric uses events.EventSliceConfigDeletionFailed (wrong).
- Read line 186 —
util.RecordEvent uses events.EventSliceQoSConfigDeleted (correct).
- Read line 190 —
mf.RecordCounterMetric uses events.EventSliceConfigDeleted (wrong).
- At runtime: delete a SliceQoSConfig and query
kubeslice_controller_events_total — counter increments under SliceConfig event names instead of QoSConfig names.
👍 Expected behavior
Prometheus counters for SliceQoSConfig deletion carry event="EventSliceQoSConfigDeletionFailed" and event="EventSliceQoSConfigDeleted" labels respectively.
👎 Actual Behavior
SliceQoSConfig deletion increments EventSliceConfigDeletionFailed and EventSliceConfigDeleted counters. SliceQoSConfig deletions are invisible in QoS-specific metric dashboards and alerts.
🐚 Relevant log output
Version
main branch — run git describe --tags in the repo
🖥️ What operating system are you seeing the problem on?
Linux
✅ Proposed Solution
At line 179, change events.EventSliceConfigDeletionFailed to events.EventSliceQoSConfigDeletionFailed. At line 190, change events.EventSliceConfigDeleted to events.EventSliceQoSConfigDeleted.
👀 Have you spent some time to check if this issue has been raised before?
Code of Conduct
📜 Description
DeleteSliceQoSConfigrecords two Prometheus counter metrics with incorrecteventlabel values. At line 179, the metric is recorded withstring(events.EventSliceConfigDeletionFailed)instead ofstring(events.EventSliceQoSConfigDeletionFailed). At line 190, it is recorded withstring(events.EventSliceConfigDeleted)instead ofstring(events.EventSliceQoSConfigDeleted). The Kubernetes Event recorder calls on the same lines correctly use the QoSConfig-specific constants. Only the Prometheuseventlabel is wrong, causing QoSConfig deletion metrics to be attributed toSliceConfigevents in all dashboards and alerts.👟 Reproduction steps
👟 Reproduction steps
service/slice_qos_config_service.go:174—util.RecordEventusesevents.EventSliceQoSConfigDeletionFailed(correct).mf.RecordCounterMetricusesevents.EventSliceConfigDeletionFailed(wrong).util.RecordEventusesevents.EventSliceQoSConfigDeleted(correct).mf.RecordCounterMetricusesevents.EventSliceConfigDeleted(wrong).kubeslice_controller_events_total— counter increments under SliceConfig event names instead of QoSConfig names.👍 Expected behavior
Prometheus counters for SliceQoSConfig deletion carry
event="EventSliceQoSConfigDeletionFailed"andevent="EventSliceQoSConfigDeleted"labels respectively.👎 Actual Behavior
SliceQoSConfig deletion increments
EventSliceConfigDeletionFailedandEventSliceConfigDeletedcounters. SliceQoSConfig deletions are invisible in QoS-specific metric dashboards and alerts.🐚 Relevant log output
Version
main branch — run
git describe --tagsin the repo🖥️ What operating system are you seeing the problem on?
Linux
✅ Proposed Solution
At line 179, change
events.EventSliceConfigDeletionFailedtoevents.EventSliceQoSConfigDeletionFailed. At line 190, changeevents.EventSliceConfigDeletedtoevents.EventSliceQoSConfigDeleted.👀 Have you spent some time to check if this issue has been raised before?
Code of Conduct