Expose it as a service.

kubernetes/autoscale/k8s/autoscale.yaml

@@ -4,7 +4,6 @@ metadata:
   name: autoscale-test
   namespace: default
 spec:
-  replicas: 3
   selector:
     matchLabels:
       app: autoscale-test
@@ -33,3 +32,16 @@ spec:
             requests:
               cpu: "250m"
               memory: "256Mi"
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: autoscale-test
+  namespace: default
+spec:
+  selector:
+    app: autoscale-test
+  type: LoadBalancer
+  ports:
+    - port: 80
+      targetPort: "http"

kubernetes/autoscale/service/gunicorn.conf.py

@@ -1,10 +1,9 @@
-import multiprocessing
 import os
 
 env_port = int(os.environ.get("PORT", 8080))
 
 bind = f"0.0.0.0:{env_port}"
-workers = max(2, multiprocessing.cpu_count() * 2 + 1)
+workers = int(os.environ.get("WORKERS", 8))
 preload_app = True
 max_requests = 100
 max_requests_jitter = 50