Update deploy.yaml

main/openshift/manifests/deploy.yaml

@@ -3,6 +3,9 @@ kind: Deployment
 metadata:
   name: phpinfo
 spec:
+  # This manifest ensures that replicas are distributed across two availability zones and placed on separate worker nodes.
+  # Deploying 4 replicas will ensure they are spread across 4 different nodes in two availability zones.
+  # Deploying 2 replicas will ensure they are spread across 2 different nodes in two availability zones.
   replicas: 4
   selector:
     matchLabels:
@@ -13,18 +16,28 @@ spec:
         app: phpinfo
     spec:
       topologySpreadConstraints:
-        - maxSkew: 1
-          topologyKey: topology.kubernetes.io/zone
-          whenUnsatisfiable: DoNotSchedule
-          labelSelector:
-            matchLabels:
-              app: phpinfo
-        - maxSkew: 1
-          topologyKey: kubernetes.io/hostname
-          whenUnsatisfiable: DoNotSchedule
-          labelSelector:
-            matchLabels:
-              app: phpinfo
+      - labelSelector:
+          matchLabels:
+            app: phpinfo
+        # maxSkew is the parameter used to evenly distribute workloads across specified topology keys.  
+        # When maxSkew is set to 1, the workload is balanced such that the difference between the highest and lowest workload across topological regions does not exceed 1 unit.
+        maxSkew: 1
+        # In this configuration, the topology key represents an availability zone, such as a distinct data center.
+        topologyKey: topology.kubernetes.io/zone
+        # If the scheduler cannot find an available worker node in the specified availability zone, the replica will not be scheduled.  
+        # This may block the deployment if no nodes are available, but it is essential to maintain high availability and resilience.  
+        whenUnsatisfiable: DoNotSchedule
+      - labelSelector:
+          matchLabels:
+            app: phpinfo
+        # maxSkew is the parameter used to evenly distribute workloads across specified topology keys.  
+        # When maxSkew is set to 1, the workload is balanced such that the difference between the highest and lowest workload across topological regions does not exceed 1 unit.
+        maxSkew: 1
+        # In this configuration, the topology key represents a hostname, indicating a specific worker node.
+        topologyKey: kubernetes.io/hostname
+        # If the scheduler cannot find an available worker node, the replica will not be scheduled.
+        # This behavior may block the deployment if no nodes are available, but it ensures high availability and resilience.
+        whenUnsatisfiable: DoNotSchedule
       containers:
       - command:
         - php
@@ -36,14 +49,23 @@ spec:
         name: phpinfo
         ports:
         - containerPort: 9000
+        # It is highly recommended to configure resources limitations to avoid any container can request too much memory or CPU from the host node
         resources:
+          # Limits are set higher than requests, effectively overcommitting resources.
+          # This approach carries some risk because if all containers simultaneously utilize resources up to their limits, the node could exceed its capacity, potentially triggering the OOM killer or causing CPU throttling.
           limits:
             cpu: 40m
-            memory: 40M
+            memory: 40M        
+          # Requests define the minimum CPU and memory resources required to schedule the replica.
+          # A replica will not be scheduled if sufficient resources are not available on the node to meet the requests.
+          # However, these resources are not reserved for the replica. The scheduler only ensures they are available at the time of scheduling.
           requests:
             cpu: 20m
             memory: 20M
         securityContext:
+          # This setting makes the container's filesystem read-only.
+          # It is considered a best practice for enhancing security.
+          # If your application requires writing to the container's filesystem, you must mount a writable temporary filesystem for that purpose.
           readOnlyRootFilesystem: true
         volumeMounts:
         - mountPath: /data