Day 4 done, frontend app is deployed in K8s

README.md

@@ -98,17 +98,19 @@ Which in turn creates the frontend Docker, yay!
 
 The GH also contains a job to push to ECR, which is not tested locally.
 
-### Day 4: Deploy the Django app in K8s
+### Day 4: Deploy the Django app in K8s using the ECR image
 
-Needless to say it's more complicated without using EKS, but let's try.
+Commit: https://github.com/lucabrunox/learning/tree/e296b828cb5
+
+Needless to say that without EKS it's more complicated, but worth the learnings.
 
 Learnings:
 - Using CronJob to get AWS creds from the node, login to ECR, and store the secret for pulling images.
 - CronJob doesn't start immediately, need to wait a minute.
 - Need to untaint control plane node in other to schedule pods.
 - Need to build the frontend image for ARM, obviously.
-- Python app fails because it can't find the UTC timezone.
-- Cannot change deployment matching labels.
+- Python app fails because it can't find the UTC timezone, needs tzdata.
+- Cannot change the matching labels of a K8s deployment.
 
 At the end we're able to execute the following kubectl on the EC2 instance to deploy the app and watch it working:
 

frontend/k8s/manifest.yaml

@@ -18,7 +18,7 @@ spec:
         - name: ecrsecret
       containers:
         - name: frontend
-          image: 658456307177.dkr.ecr.eu-west-1.amazonaws.com/learning-frontend:vfca136021ea7fd89ad2bd60e47fe275d694eafaa
+          image: 658456307177.dkr.ecr.eu-west-1.amazonaws.com/learning-frontend:ve296b828cb5109a608881efa9fe5bf208d682bbc
           imagePullPolicy: IfNotPresent
           ports:
             - containerPort: 8000