Add startup probe to determine if model is downloaded

Signed-off-by: Jonathan Minkin <minkinj@amazon.com>

helm-charts/common/ollama-service/templates/deployment.yaml

@@ -68,10 +68,19 @@ spec:
             - name: http
               containerPort: {{ .Values.service.targetPort }}
               protocol: TCP
+          {{- if .Values.livenessProbe }}
           livenessProbe:
             {{- toYaml .Values.livenessProbe | nindent 12 }}
+          {{- end }}
+          {{- if .Values.readinessProbe }}
           readinessProbe:
             {{- toYaml .Values.readinessProbe | nindent 12 }}
+          {{- end }}
+          startupProbe:
+            httpGet:
+              # Ollama uses lazy loading for models, but this verifies the model has been downloaded
+              path: /v1/models/{{ .Values.LLM_MODEL_ID }}
+              port: http
           resources:
             {{- toYaml .Values.resources | nindent 12 }}
           volumeMounts:

helm-charts/common/ollama-service/values.yaml

@@ -77,14 +77,14 @@ resources: {}
   #   memory: 128Mi
 
 # This is to setup the liveness and readiness probes more information can be found here: https://kubernetes.io/docs/tasks/configure-pod-container/configure-liveness-readiness-startup-probes/
-livenessProbe:
-  httpGet:
-    path: /api/version
-    port: http
-readinessProbe:
-  httpGet:
-    path: /api/version
-    port: http
+# livenessProbe:
+#   httpGet:
+#     path: /api/version
+#     port: http
+# readinessProbe:
+#   httpGet:
+#     path: /api/version
+#     port: http
 
 
 LLM_MODEL_ID: "llama3.2:1b"