From b64bfcc414ec4faddecf49ff9ece8ef169896bc0 Mon Sep 17 00:00:00 2001 From: Arsolitt Date: Sat, 18 Apr 2026 07:36:14 +0300 Subject: [PATCH] fix(dashboard): deduplicate pending GPU pods by (namespace, pod) The Pending GPU pods counter on gpu-quotas joined raw kube_pod_container_resource_requests (per-container series) against kube_pod_status_phase (per-pod series). Multi-container pods were counted once per requesting container instead of once per pod, so the widget over-reported whenever a Pending pod had more than one GPU container. Collapse the requests to pod level before the join. Signed-off-by: Arsolitt --- dashboards/gpu/gpu-quotas.json | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/dashboards/gpu/gpu-quotas.json b/dashboards/gpu/gpu-quotas.json index 3bc01a83..624f072b 100644 --- a/dashboards/gpu/gpu-quotas.json +++ b/dashboards/gpu/gpu-quotas.json @@ -208,7 +208,7 @@ "id": 5, "targets": [ { - "expr": "count((kube_pod_container_resource_requests{resource=\"nvidia_com_gpu\"} \u003e 0) * on(namespace, pod) group_left() (kube_pod_status_phase{phase=\"Pending\"} == 1)) or vector(0)", + "expr": "count((sum by (namespace, pod) (kube_pod_container_resource_requests{resource=\"nvidia_com_gpu\"}) \u003e 0) * on(namespace, pod) group_left() (kube_pod_status_phase{phase=\"Pending\"} == 1)) or vector(0)", "refId": "A" } ],