diff --git a/REVIEW.md b/REVIEW.md
new file mode 100644
index 0000000..9ec2782
--- /dev/null
+++ b/REVIEW.md
@@ -0,0 +1,46 @@
+# Review Guide
+
+This doc is mainly for reviewers.
+
+For testing, a Chaos Mesh instance is required. In next section, we will show how to deploy a Chaos Mesh instance with Helm.
+
+## Deploy Chaos Mesh
+
+> Refer to <https://chaos-mesh.org/docs/production-installation-using-helm/> for more details.
+
+Follow the steps below to deploy a minimal Chaos Mesh instance:
+
+```bash
+# Add the chart
+helm repo add chaos-mesh https://charts.chaos-mesh.org
+# Create a Chaos Mesh instance
+# Refer to https://chaos-mesh.org/docs/production-installation-using-helm/#step-4-install-chaos-mesh-in-different-environments if you use containerd or other container runtimes.
+helm install chaos-mesh chaos-mesh/chaos-mesh -n=chaos-mesh --create-namespace --set controllerManager.leaderElection.enabled=false,dashboard.securityMode=false
+```
+
+Verify the installation:
+
+```bash
+kubectl get pods -n chaos-mesh
+```
+
+Port-forward the dashboard:
+
+```bash
+kubectl port-forward svc/chaos-dashboard -n chaos-mesh 2333:2333
+```
+
+Then you can visit the dashboard at <http://localhost:2333>.
+
+## Test data source
+
+For data source testing, there must be some Chaos events are already created. You can finish this step by creating a Chaos experiment in the dashboard.
+
+Navigate to `Experiments` at the left sidebar and click `New experiment` button. For example,
+let's create a `Pod Failure` experiment. Select `Kubernetes` -> `Pod Fault` -> `Pod Failure`, then
+choose `Namespace Selectors` in `Scope` to `default` (this means the experiment will be applied to
+all pods in the `default` namespace). Finally, fill in a name like `pod-failure` and set `Duration` to `1m`,
+and click `Submit` button to create the experiment.
+
+Now you can test the data source. Start creating a query from here: <https://github.com/chaos-mesh/datasource?tab=readme-ov-file#query>.
+At the same time, you can also view created Chaos events via <http://localhost:2333/#/events>.