diff --git a/README.md b/README.md
index 77bdf06d..629d2160 100644
--- a/README.md
+++ b/README.md
@@ -177,6 +177,10 @@ Kubernetes cluster to a healthy state. The following remedy systems exist:
   [this issue](https://github.com/kubernetes/node-problem-detector/issues/199)
   for an example production use case for Draino.
 
+# Docs
+
+* [Custom plugin monitor](docs/custom_plugin_monitor.md)
+
 # Links
 
 * [Design Doc](https://docs.google.com/document/d/1cs1kqLziG-Ww145yN6vvlKguPbQQ0psrSBnEqpy0pzE/edit?usp=sharing)
diff --git a/docs/custom_plugin_monitor.md b/docs/custom_plugin_monitor.md
new file mode 100644
index 00000000..b8ad325e
--- /dev/null
+++ b/docs/custom_plugin_monitor.md
@@ -0,0 +1,8 @@
+# Custom Plugin Monitor
+
+## Configuration
+### Plugin Config
+* `invoke_interval`: Interval at which custom plugins will be invoked.
+* `timeout`: Time after which custom plugins invokation will be terminated and considered timeout.
+* `max_output_length`: The maximum standard output size from custom plugins that NPD will be cut and use for condition status message.
+* `concurrency`: The plugin worker number, i.e., how many custom plugins will be invoked concurrently.