ContextYou are asked to allow a Pod to communicate with two other Pods but nothing...

ssh ckad00018

You cannot create/modify/delete any NetworkPolicy.

So the only way to make the existing policies “take effect” is to ensure the right Pods have the labels/selectors those policies expect.

The task: in namespace charming-macaw, configure things so the target Pod can send + receive traffic ONLY to/from Pods front and db.

1) Inspect what NetworkPolicies already exist (don’t change them)

kubectl -n charming-macaw get netpol

kubectl -n charming-macaw get netpol -o wide

Dump them to see the selectors they use:

kubectl -n charming-macaw get netpol -o yaml

You are looking for policies that:

select the restricted pod via spec.podSelector

and allow ingress/egress only with selectors that match front and db

often there’s also a “default deny” policy.

2) Identify the Pods and their current labels

kubectl -n charming-macaw get pods -o wide

kubectl -n charming-macaw get pods --show-labels

Specifically inspect labels for front and db:

kubectl -n charming-macaw get pod front --show-labels

kubectl -n charming-macaw get pod db --show-labels

(If they’re Deployments instead of single Pods, do:)

kubectl -n charming-macaw get deploy --show-labels

kubectl -n charming-macaw get pods -l app=front --show-labels

kubectl -n charming-macaw get pods -l app=db --show-labels

3) Figure out which pod is “the Pod” to restrict

Usually there’s a third pod (e.g., backend, api, app) besides front and db.

List pods again and identify the “other” one:

kubectl -n charming-macaw get pods

Let’s assume the pod to restrict is called app (replace as needed):

TARGET=

4) Match the existing NetworkPolicy selectors by labeling pods (allowed)

Because you can’t edit NetworkPolicies, you must make labels on Pods (or their controllers) match the policies’ selectors.

4.1 Determine the label required on the TARGET pod

From the YAML, find the policy that selects the restricted pod, e.g.:

spec:

podSelector:

matchLabels:

role: restricted

Extract podSelector from each policy quickly:

kubectl -n charming-macaw get netpol -o jsonpath='{range .items[*]}{.metadata.name}{" => "}{.spec.podSelector}{"\n"}{end}'

Pick the selector that is meant for the restricted pod, then apply it to the TARGET pod (example: role=restricted):

kubectl -n charming-macaw label pod $TARGET role=restricted --overwrite

Best practice (if the pod is managed by a Deployment): label the Deployment template instead, so it persists.

Find the owner:

kubectl -n charming-macaw get pod $TARGET -o jsonpath='{.metadata.ownerReferences[0].kind}{" "}{.metadata.ownerReferences[0].name}{"\n"}'

If it’s a ReplicaSet, find its Deployment:

RS=$(kubectl -n charming-macaw get pod $TARGET -o jsonpath='{.metadata.ownerReferences[0].name}')

kubectl -n charming-macaw get rs $RS -o jsonpath='{.metadata.ownerReferences[0].kind}{" "}{.metadata.ownerReferences[0].name}{"\n"}'

Then label the Deployment (example):

kubectl -n charming-macaw label deploy role=restricted --overwrite

4.2 Ensure front and db match what the allow-rules reference

Look inside the allow policy ingress.from / egress.to. You might see something like:

from:

- podSelector:

matchLabels:

name: front

- podSelector:

matchLabels:

name: db

So you must ensure:

front pod has name=front

db pod has name=db

Apply labels (examples—use what the policy expects):

kubectl -n charming-macaw label pod front name=front --overwrite

kubectl -n charming-macaw label pod db name=db --overwrite

Again, if they’re Deployments, label the Deployment instead:

kubectl -n charming-macaw label deploy front name=front --overwrite

kubectl -n charming-macaw label deploy db name=db --overwrite

5) Verify the NetworkPolicies now “select” the right pods

Check which labels each pod has now:

kubectl -n charming-macaw get pods --show-labels

Confirm the restricted pod matches the NetPol podSelector:

kubectl -n charming-macaw get netpol -o jsonpath='{.spec.podSelector}{"\n"}'

kubectl -n charming-macaw get pod $TARGET --show-labels

6) Functional verification (quick network tests)

Exec into the restricted pod and try to reach:

front ✅ allowed

db ✅ allowed

anything else ❌ blocked

If busybox has wget:

kubectl -n charming-macaw exec -it $TARGET -- sh -c 'wget -qO- http://front 2 >/dev/null || true'

kubectl -n charming-macaw exec -it $TARGET -- sh -c 'wget -qO- http://db 2 >/dev/null || true'

Test something that should be blocked (example: kubernetes service DNS name):

kubectl -n charming-macaw exec -it $TARGET -- sh -c 'wget -qO- https://kubernetes.default.svc 2 >/dev/null || echo "blocked"'

Also test inbound (from front to target, and from db to target) if the target listens on a port; otherwise inbound testing may be limited.

What you’re doing conceptually

Existing NetPols are already correct.

Your job is to make pod labels match the NetPol selectors so:

default deny applies to the target

allow rules apply only between target ↔ front and target ↔ db