feat: add startup probe for machine learning

[samholton]

I ran into issues with machine-learning pod not having enough time to download and startup after adding MACHINE_LEARNING_PRELOAD__CLIP.

[github-actions]

Changes in Final Manifest

• manifest.yaml

manifest.yaml

--- a/main/charts/immich/build/manifest.yaml
+++ b/charts/immich/build/manifest.yaml
@@ -10,7 +10,7 @@ metadata:
     app.kubernetes.io/managed-by: Helm
     app.kubernetes.io/name: machine-learning
     app.kubernetes.io/version: v1.119.0
-    helm.sh/chart: immich-0.8.4
+    helm.sh/chart: immich-0.8.5
   annotations:
 spec:
   type: ClusterIP
@@ -34,7 +34,7 @@ metadata:
     app.kubernetes.io/managed-by: Helm
     app.kubernetes.io/name: server
     app.kubernetes.io/version: v1.119.0
-    helm.sh/chart: immich-0.8.4
+    helm.sh/chart: immich-0.8.5
   annotations:
 spec:
   type: ClusterIP
@@ -57,7 +57,7 @@ metadata:
     app.kubernetes.io/managed-by: Helm
     app.kubernetes.io/name: machine-learning
     app.kubernetes.io/version: v1.119.0
-    helm.sh/chart: immich-0.8.4
+    helm.sh/chart: immich-0.8.5
 spec:
   revisionHistoryLimit: 3
   replicas: 1
@@ -120,6 +120,14 @@ spec:
             initialDelaySeconds: 0
             periodSeconds: 10
             timeoutSeconds: 1
+          startupProbe:
+            failureThreshold: 60
+            httpGet:
+              path: /ping
+              port: http
+            initialDelaySeconds: 0
+            periodSeconds: 10
+            timeoutSeconds: 1
       volumes:
         - name: cache
           emptyDir:
@@ -135,7 +143,7 @@ metadata:
     app.kubernetes.io/managed-by: Helm
     app.kubernetes.io/name: server
     app.kubernetes.io/version: v1.119.0
-    helm.sh/chart: immich-0.8.4
+    helm.sh/chart: immich-0.8.5
 spec:
   revisionHistoryLimit: 3
   replicas: 1

See the workflow run