diff --git a/metrics/deploy/prometheus-ocs-rules.yaml b/metrics/deploy/prometheus-ocs-rules.yaml
index 2bbfe59209..0a5d11745e 100644
--- a/metrics/deploy/prometheus-ocs-rules.yaml
+++ b/metrics/deploy/prometheus-ocs-rules.yaml
@@ -282,7 +282,7 @@ spec:
         message: High MDS cache usage for the daemon {{ $labels.ceph_daemon }}.
         severity_level: error
       expr: |
-        ceph_mds_mem_rss / on(ceph_daemon) group_left(job)(label_replace(kube_pod_container_resource_requests{container="mds", resource="memory"}, "ceph_daemon", "mds.$1", "pod", "rook-ceph-mds-(.*)-(.*)") * .5) > .95
+        (ceph_mds_mem_rss * 1000) / on(ceph_daemon) group_left(job)(label_replace(kube_pod_container_resource_requests{container="mds", resource="memory"}, "ceph_daemon", "mds.$1", "pod", "rook-ceph-mds-(.*)-(.*)") * .5) > .95
       for: 5m
       labels:
         severity: critical
diff --git a/metrics/mixin/alerts/perf.libsonnet b/metrics/mixin/alerts/perf.libsonnet
index ae385de236..0cc0ea4b18 100644
--- a/metrics/mixin/alerts/perf.libsonnet
+++ b/metrics/mixin/alerts/perf.libsonnet
@@ -7,7 +7,7 @@
           {
             alert: 'MDSCacheUsageHigh',
             expr: |||
-              ceph_mds_mem_rss / on(ceph_daemon) group_left(job)(label_replace(kube_pod_container_resource_requests{container="mds", resource="memory"}, "ceph_daemon", "mds.$1", "pod", "rook-ceph-mds-(.*)-(.*)") * .5) > .95
+              (ceph_mds_mem_rss * 1000) / on(ceph_daemon) group_left(job)(label_replace(kube_pod_container_resource_requests{container="mds", resource="memory"}, "ceph_daemon", "mds.$1", "pod", "rook-ceph-mds-(.*)-(.*)") * .5) > .95
             ||| % $._config,
             'for': $._config.mdsCacheUsageAlertTime,
             labels: {