diff --git a/docs/site/docs/guides/kubernetes-audit-logs.md b/docs/site/docs/guides/kubernetes-audit-logs.md
new file mode 100644
index 00000000..a9d8b5e7
--- /dev/null
+++ b/docs/site/docs/guides/kubernetes-audit-logs.md
@@ -0,0 +1,73 @@
+---
+title: Analyze Kubernetes Audit Logs
+description: Query Kubernetes API audit JSON Lines with LynxDB to find anonymous requests, source IPs, and targeted resources.
+---
+
+# Analyze Kubernetes Audit Logs
+
+The Kubernetes API server's log audit backend writes one JSON event per line.
+When the backend uses its common `/var/log/kubernetes/audit/audit.log` path,
+LynxDB can analyze the file directly in pipe mode without starting a server or
+ingesting the data first.
+
+See the Kubernetes documentation for [configuring the audit log
+backend](https://kubernetes.io/docs/tasks/debug/debug-cluster/audit/#log-backend)
+and the [audit Event field
+reference](https://kubernetes.io/docs/reference/config-api/apiserver-audit.v1/#audit-k8s-io-v1-Event).
+Kubernetes records an [anonymous
+request](https://kubernetes.io/docs/reference/access-authn-authz/authentication/#anonymous-requests)
+with the username `system:anonymous` and membership in the
+`system:unauthenticated` group.
+
+## Find anonymous API requests
+
+This query groups anonymous requests by user agent and summarizes their source
+IPs and targeted resource types:
+
+```bash
+sudo cat /var/log/kubernetes/audit/audit.log | lynxdb query '
+  | parse json
+  | where user.username == "system:anonymous"
+  | extend source_ips_array = from_json(sourceIPs)
+  | explode source_ips_array as source_ip
+  | stats dc(auditID) as requests,
+          values(source_ip, 20) as source_ips,
+          values(objectRef.resource, 20) as resources
+    by userAgent
+  | sort -requests'
+```
+
+The pipeline does the following:
+
+1. `parse json` exposes nested fields such as `user.username` and
+   `objectRef.resource`.
+2. `from_json(sourceIPs)` restores the JSON-encoded IP list, and `explode`
+   emits one row per address.
+3. `dc(auditID)` counts unique API requests after that expansion, so requests
+   with multiple proxy addresses are not counted more than once.
+4. `values(..., 20)` keeps each summary bounded to 20 distinct IPs and resource
+   types per user agent.
+
+:::caution
+Kubernetes documents `auditID` as a unique ID for each request. It also warns
+that proxy entries in `sourceIPs` can be supplied by the client and that
+`userAgent` is untrusted client input. Treat both fields as investigation clues,
+not identity proof.
+:::
+
+## Summarize failed requests
+
+To see which API operations and resources produce error responses, aggregate
+the response code instead:
+
+```bash
+sudo cat /var/log/kubernetes/audit/audit.log | lynxdb query '
+  | parse json
+  | where responseStatus.code >= 400
+  | stats count() as requests
+    by responseStatus.code, verb, objectRef.resource
+  | sort -requests'
+```
+
+`objectRef` is absent from some non-resource and list requests, so those groups
+may have a null resource value.
diff --git a/docs/site/sidebars.ts b/docs/site/sidebars.ts
index b5b65af2..9651ae6d 100644
--- a/docs/site/sidebars.ts
+++ b/docs/site/sidebars.ts
@@ -35,6 +35,7 @@ const sidebars: SidebarsConfig = {
         'guides/aggregations',
         'guides/field-extraction',
         'guides/json-processing',
+        'guides/kubernetes-audit-logs',
         'guides/time-series',
         'guides/materialized-views',
         'guides/saved-queries',