bump prometheus memory

running into OOM during db reconstruction

and reconstruction takes too long for liveness probe

Author: minrk

config/prod.yaml

@@ -17,7 +17,7 @@ binderhub:
       sticky_builds: true
 
   nodeSelector: *coreNodeSelector
-  
+
   resources:
     requests:
       cpu: 0.5
@@ -112,13 +112,14 @@ grafana:
 prometheus:
   server:
     nodeSelector: *coreNodeSelector
+    livenessProbeInitialDelay: 600
     resources:
       requests:
         cpu: "2"
-        memory: 16Gi
+        memory: 24Gi
       limits:
         cpu: "2"
-        memory: 16Gi
+        memory: 24Gi
     persistentVolume:
       # Use a large SSD Volume in production
       size: 2000Gi