diff --git a/chart/templates/backend.yaml b/chart/templates/backend.yaml
index dd148d63..ea9c7441 100644
--- a/chart/templates/backend.yaml
+++ b/chart/templates/backend.yaml
@@ -10,7 +10,9 @@ spec:
     matchLabels:
       app: {{ .Values.name }}
       role: backend
-  replicas: {{ .Values.backend_num_replicas }}
+{{- if eq (int .Values.backend_max_replicas) 1 }}
+  replicas: 1
+{{- end }}
   template:
     metadata:
       labels:
@@ -232,3 +234,31 @@ spec:
     - protocol: TCP
       port: {{ .Values.opPort }}
       name: operator
+
+{{- if gt (int .Values.backend_max_replicas) 1 }}
+---
+apiVersion: autoscaling/v2
+kind: HorizontalPodAutoscaler
+metadata:
+  name: backend-autoscaler
+spec:
+  scaleTargetRef:
+    apiVersion: apps/v1
+    kind: Deployment
+    name: {{ .Values.name }}-backend
+  minReplicas: 1
+  maxReplicas: {{ .Values.backend_max_replicas }}
+  metrics:
+  - type: Resource
+    resource:
+      name: cpu
+      target:
+        type: Utilization
+        averageUtilization: {{ .Values.backend_avg_cpu_threshold }}
+  - type: Resource
+    resource:
+      name: memory
+      target:
+        type: Utilization
+        averageUtilization: {{ .Values.backend_avg_memory_threshold }}
+{{- end }}
diff --git a/chart/templates/frontend.yaml b/chart/templates/frontend.yaml
index f3a3d39a..45278586 100644
--- a/chart/templates/frontend.yaml
+++ b/chart/templates/frontend.yaml
@@ -10,7 +10,9 @@ spec:
     matchLabels:
       app: {{ .Values.name }}
       role: frontend
-  replicas: {{ .Values.frontend_num_replicas | default 1 }}
+  {{- if eq (int .Values.frontend_max_replicas) 1 }}
+  replicas: 1
+  {{- end }}
   template:
     metadata:
       labels:
@@ -110,4 +112,30 @@ spec:
   {{- end }}
 
 
-
+{{- if gt (int .Values.frontend_max_replicas) 1 }}
+---
+apiVersion: autoscaling/v2
+kind: HorizontalPodAutoscaler
+metadata:
+  name: frontend-autoscaler
+spec:
+  scaleTargetRef:
+    apiVersion: apps/v1
+    kind: Deployment
+    name: {{ .Values.name }}-frontend
+  minReplicas: 1
+  maxReplicas: {{ .Values.frontend_max_replicas }}
+  metrics:
+  - type: Resource
+    resource:
+      name: cpu
+      target:
+        type: Utilization
+        averageUtilization: {{ .Values.frontend_avg_cpu_threshold }}
+  - type: Resource
+    resource:
+      name: memory
+      target:
+        type: Utilization
+        averageUtilization: {{ .Values.frontend_avg_memory_threshold }}
+{{ end }}
diff --git a/chart/values.yaml b/chart/values.yaml
index cbae7fb4..f2cea90d 100644
--- a/chart/values.yaml
+++ b/chart/values.yaml
@@ -91,13 +91,10 @@ backend_pull_policy: "Always"
 
 backend_password_secret: "PASSWORD!"
 
-# number of backend pods
-backend_num_replicas: 1
-
 # number of workers per pod
 backend_workers: 1
 
-backend_cpu: "25m"
+backend_cpu: "100m"
 
 backend_memory: "350Mi"
 
@@ -114,6 +111,17 @@ profile_browser_idle_seconds: 60
 # mostly intended for debugging / testing
 # log_failed_crawl_lines: 200
 
+# Autoscale
+# ---------
+# max number of backend pods to scale to
+# if > 1, will enable HPA for backend
+backend_max_replicas: 1
+
+# scale up if avg cpu utilization exceeds
+backend_avg_cpu_threshold: 80
+
+# scale up if avg memory utilization exceeds
+backend_avg_memory_threshold: 95
 
 # Nginx Image
 # =========================================
@@ -132,6 +140,18 @@ local_service_port: 30870
 
 frontend_alias: "http://browsertrix-cloud-frontend"
 
+# Autoscaling
+# -----------
+# max number of backend pods to scale to
+# if > 1, will enable HPA for frontend
+frontend_max_replicas: 1
+
+# scale up if avg cpu utilization exceeds
+frontend_avg_cpu_threshold: 80
+
+# scale up if avg memory utilization exceeds
+frontend_avg_memory_threshold: 95
+
 
 # MongoDB Image
 # =========================================
diff --git a/docs/deploy/customization.md b/docs/deploy/customization.md
index f145ff7f..e50349e9 100644
--- a/docs/deploy/customization.md
+++ b/docs/deploy/customization.md
@@ -78,6 +78,36 @@ storages:
     endpoint_url: "http://s3provider.example.com"
 ```
 
+## Horizontal Autoscaling
+
+Browsertrix also includes support for horizontal auto-scaling for both the backend and frontend pods.
+The auto-scaling will start a new pod when memory/cpu utilization reaches the thresholds.
+
+To use auto-scaling, the [metrics-server](https://github.com/kubernetes-sigs/metrics-server) cluster add-on is required.
+Many k8s provides include metrics server by default, others, like MicroK8S, make it available as an add-on.
+
+To enable auto-scaling, set `backend_max_replicas` and/or `frontend_max_replicas` to a value >1.
+
+```yaml
+backend_max_replicas: 2
+
+frontend_max_replicas: 2
+```
+
+By default, the auto-scaling uses the following thresholds for deciding when to start a new pod can also
+be modified. The default values are:
+
+```yaml
+backend_avg_cpu_threshold: 80
+
+backend_avg_memory_threshold: 95
+
+frontend_avg_cpu_threshold: 80
+
+frontend_avg_memory_threshold: 95
+```
+
+
 ## Email / SMTP Server
 
 Browsertrix sends user invitations, password resets, background job failure notifications, and other important messages via email. The `email` setting can be used to configure the SMTP server used to send emails. To avoid email messages from Browsertrix being flagged as spam, be sure to use the same domain for `sender_email` and `reply_to_email`.