diff --git a/CHANGELOG.md b/CHANGELOG.md
index b4ca603..a37ba0b 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,6 +7,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Added
+- Added configurable deployment strategy support, allowing users to switch between `RollingUpdate` and `Recreate` strategies. [#82](https://github.com/sourcebot-dev/sourcebot-helm-chart/pull/82)
+
 ## [0.1.71] - 2026-04-09
 
 ### Added
diff --git a/README.md b/README.md
index 7addfc1..bc1dde3 100644
--- a/README.md
+++ b/README.md
@@ -221,6 +221,37 @@ sourcebot:
     existingSecretKey: license-key
 ```
 
+## Deployment Strategy
+
+By default, the chart uses a `RollingUpdate` strategy, which creates the new pod before terminating the old one during upgrades. You can customize the rolling update behavior:
+```yaml
+sourcebot:
+  strategy:
+    type: RollingUpdate
+    rollingUpdate:
+      maxUnavailable: 0
+      maxSurge: 1
+```
+
+On multi-node clusters with `ReadWriteOnce` persistent volumes, the new pod may fail to start if it gets scheduled on a different node than the old pod. To avoid this, you can pin pods to the same node using pod affinity:
+```yaml
+sourcebot:
+  affinity:
+    podAffinity:
+      requiredDuringSchedulingIgnoredDuringExecution:
+        - labelSelector:
+            matchLabels:
+              app.kubernetes.io/name: sourcebot
+          topologyKey: kubernetes.io/hostname
+```
+
+Alternatively, you can use the `Recreate` strategy, which terminates the old pod before creating the new one. This avoids PVC conflicts but causes brief downtime during upgrades:
+```yaml
+sourcebot:
+  strategy:
+    type: Recreate
+```
+
 ## Persistence
 
 Each component has its own persistent volume for storing data across pod restarts.
@@ -325,6 +356,31 @@ sourcebot:
         secretName: sourcebot-tls
 ```
 
+### Zero-downtime upgrades on multi-node clusters
+
+Use a `ReadWriteMany` access mode so the old and new pods can mount the PVC simultaneously during rolling updates. This requires a storage class that supports `ReadWriteMany` (e.g., EFS on AWS, Filestore on GCP, Azure Files):
+
+```yaml
+sourcebot:
+  persistence:
+    accessModes:
+      - ReadWriteMany
+    storageClass: "efs-sc"  # example: AWS EFS storage class
+```
+
+Alternatively, you can keep `ReadWriteOnce` and use pod affinity to pin pods to the same node:
+
+```yaml
+sourcebot:
+  affinity:
+    podAffinity:
+      requiredDuringSchedulingIgnoredDuringExecution:
+        - labelSelector:
+            matchLabels:
+              app.kubernetes.io/name: sourcebot
+          topologyKey: kubernetes.io/hostname
+```
+
 ### Using an existing PVC
 
 ```yaml
diff --git a/charts/sourcebot/README.md b/charts/sourcebot/README.md
index bac9c55..41c3842 100644
--- a/charts/sourcebot/README.md
+++ b/charts/sourcebot/README.md
@@ -130,6 +130,7 @@ Sourcebot is a self-hosted tool that helps you understand your codebase.
 | sourcebot.startupProbe.httpGet.path | string | `"/api/health"` | Path to check |
 | sourcebot.startupProbe.httpGet.port | string | `"http"` | Port to check |
 | sourcebot.startupProbe.periodSeconds | int | `30` | Initial delay before the first probe |
+| sourcebot.strategy | object | `{"type":"RollingUpdate"}` | Deployment strategy configuration |
 | sourcebot.tolerations | list | `[]` | Set tolerations for pod scheduling See: https://kubernetes.io/docs/concepts/scheduling-eviction/taint-and-toleration/ |
 
 ----------------------------------------------
diff --git a/charts/sourcebot/templates/deployment.yaml b/charts/sourcebot/templates/deployment.yaml
index d64da76..7dc6d2b 100644
--- a/charts/sourcebot/templates/deployment.yaml
+++ b/charts/sourcebot/templates/deployment.yaml
@@ -8,6 +8,10 @@ metadata:
     {{- include "sourcebot.labels" $ | nindent 4 }}
 spec:
   replicas: {{ $.Values.sourcebot.replicaCount }}
+  {{- with $.Values.sourcebot.strategy }}
+  strategy:
+    {{- toYaml . | nindent 4 }}
+  {{- end }}
   selector:
     matchLabels:
       {{- include "sourcebot.selectorLabels" $ | nindent 6 }}
diff --git a/charts/sourcebot/values.schema.json b/charts/sourcebot/values.schema.json
index 1fea3fd..d6a5135 100644
--- a/charts/sourcebot/values.schema.json
+++ b/charts/sourcebot/values.schema.json
@@ -36,6 +36,26 @@
           "type": "integer",
           "minimum": 1
         },
+        "strategy": {
+          "type": "object",
+          "oneOf": [
+            {
+              "additionalProperties": false,
+              "properties": {
+                "type": { "type": "string", "const": "RollingUpdate" },
+                "rollingUpdate": { "type": "object" }
+              },
+              "required": ["type"]
+            },
+            {
+              "additionalProperties": false,
+              "properties": {
+                "type": { "type": "string", "const": "Recreate" }
+              },
+              "required": ["type"]
+            }
+          ]
+        },
         "image": {
           "type": "object",
           "properties": {
diff --git a/charts/sourcebot/values.yaml b/charts/sourcebot/values.yaml
index aa35132..80229cd 100644
--- a/charts/sourcebot/values.yaml
+++ b/charts/sourcebot/values.yaml
@@ -4,7 +4,7 @@ global:
   security:
     # -- Allow insecure images to use bitnami legacy repository. Can be set to false if secure images are being used (Paid).
     allowInsecureImages: true
-  
+
   # -- Global Docker registry secret names as an array
   imagePullSecrets: []
 
@@ -19,6 +19,10 @@ sourcebot:
   # -- Set the number of replicas for the deployment
   replicaCount: 1
 
+  # -- Deployment strategy configuration
+  strategy:
+    type: RollingUpdate
+
   # Image configuration
   image:
     # -- Container image repository