diff --git a/operator/CHANGELOG.md b/operator/CHANGELOG.md
index f0bf7f4d5b5f..49f72d84858b 100644
--- a/operator/CHANGELOG.md
+++ b/operator/CHANGELOG.md
@@ -1,5 +1,9 @@
 ## Main
 
+## Release 5.8.11
+
+- [13512](https://github.com/grafana/loki/pull/13512) **xperimental**: feat(operator): Add alert for discarded samples
+
 ## Release 5.8.10
 
 - [13562](https://github.com/grafana/loki/pull/13562) **xperimental**: fix(operator): Set object storage for delete requests when using retention
diff --git a/operator/docs/lokistack/sop.md b/operator/docs/lokistack/sop.md
index 8c437bd53b67..00ac13a071be 100644
--- a/operator/docs/lokistack/sop.md
+++ b/operator/docs/lokistack/sop.md
@@ -308,3 +308,35 @@ The query queue is currently under high load.
 ### Steps
 
 - Increase the number of queriers
+
+## Loki Discarded Samples Warning
+
+### Impact
+
+Loki is discarding samples (log entries) because they fail validation. This alert only fires for errors that are not retryable. This means that the discarded samples are lost.
+
+### Summary
+
+Loki can reject log entries (samples) during submission when they fail validation. This happens on a per-stream basis, so only the specific samples or streams failing validation are lost.
+
+The possible validation errors are documented in the [Loki documentation](https://grafana.com/docs/loki/latest/operations/request-validation-rate-limits/#validation-errors). This alert only fires for the validation errors that are not retryable, which means that discarded samples are permanently lost.
+
+The alerting can only show the affected Loki tenant. Since Loki 3.1.0 more detailed information about the affected streams is provided in an error message emitted by the distributor component.
+
+This information can be used to pinpoint the application sending the offending logs. For some of the validations there are configuration parameters that can be tuned in LokiStack's `limits` structure, if the messages should be accepted. Usually it is recommended to fix the issue either on the emitting application (if possible) or by changing collector configuration to fix non-compliant messages before sending them to Loki.
+
+### Severity
+
+`Warning`
+
+### Access Required
+
+- Console access to the cluster
+- View access in the namespace where the LokiStack is deployed
+  - OpenShift
+    - `openshift-logging` (LokiStack)
+
+### Steps
+
+- View detailed log output from the Loki distributors to identify affected streams
+- Decide on further steps depending on log source and validation error
diff --git a/operator/internal/manifests/internal/alerts/prometheus-alerts.yaml b/operator/internal/manifests/internal/alerts/prometheus-alerts.yaml
index f378c49fd78c..e0c49d61fe0a 100644
--- a/operator/internal/manifests/internal/alerts/prometheus-alerts.yaml
+++ b/operator/internal/manifests/internal/alerts/prometheus-alerts.yaml
@@ -175,3 +175,21 @@ groups:
     for: 15m
     labels:
       severity: warning
+  - alert: LokiDiscardedSamplesWarning
+    annotations:
+      message: |-
+        Loki in namespace {{ $labels.namespace }} is discarding samples in the "{{ $labels.tenant }}" tenant during ingestion.
+        Samples are discarded because of "{{ $labels.reason }}" at a rate of {{ .Value | humanize }} samples per second.
+      summary: Loki is discarding samples during ingestion because they fail validation.
+      runbook_url: "[[ .RunbookURL]]#Loki-Discarded-Samples-Warning"
+    expr: |
+      sum by(namespace, tenant, reason) (
+        irate(loki_discarded_samples_total{
+          reason!="rate_limited",
+          reason!="per_stream_rate_limit",
+          reason!="stream_limit"}[2m])
+      )
+      > 0
+    for: 15m
+    labels:
+      severity: warning
diff --git a/operator/internal/manifests/internal/alerts/testdata/test.yaml b/operator/internal/manifests/internal/alerts/testdata/test.yaml
index a4d8bec8a6a4..d60e3befa002 100644
--- a/operator/internal/manifests/internal/alerts/testdata/test.yaml
+++ b/operator/internal/manifests/internal/alerts/testdata/test.yaml
@@ -63,6 +63,9 @@ tests:
       - series: 'loki_logql_querystats_latency_seconds_bucket{namespace="my-ns", job="querier", route="my-route", le="+Inf"}'
         values: '0+100x20'
 
+      - series: 'loki_discarded_samples_total{namespace="my-ns", tenant="application", reason="line_too_long"}'
+        values: '0x5 0+120x25 3000'
+
     alert_rule_test:
       - eval_time: 16m
         alertname: LokiRequestErrors
@@ -177,3 +180,17 @@ tests:
               summary: "The read path has high volume of queries, causing longer response times."
               message: "The read path is experiencing high load."
               runbook_url: "[[ .RunbookURL ]]#Loki-Read-Path-High-Load"
+      - eval_time: 22m
+        alertname: LokiDiscardedSamplesWarning
+        exp_alerts:
+          - exp_labels:
+              namespace: my-ns
+              tenant: application
+              severity: warning
+              reason: line_too_long
+            exp_annotations:
+              message: |-
+                Loki in namespace my-ns is discarding samples in the "application" tenant during ingestion.
+                Samples are discarded because of "line_too_long" at a rate of 2 samples per second.
+              summary: Loki is discarding samples during ingestion because they fail validation.
+              runbook_url: "[[ .RunbookURL]]#Loki-Discarded-Samples-Warning"