Appendix A: Cheat Sheet
Tested with Kubernetes 1.37 (kind 0.33), Cilium 1.20, Kyverno 1.19.1
(chart 3.9.1), Kyverno CLI 1.19.1 and Chainsaw 0.2.15. Names and labels
match the demo app.
Tools in the lab
alias cc=~/guard-lab/code/demo/conncheck.sh
# A connection from inside a Pod (source: Pod name or value of app=)
cc shop/frontend http://api.shop
cc shop/api tcp://db.shop:5432
# With an expectation: exit 0 = as expected, 1 = deviation
cc default/client http://api.shop deny
# Longer timeout under load or for targets on the internet
TIMEOUT=5 cc default/client https://example.com
# Target matrix vs. actual (file: "<ns>/<pod|app> <url> allow|deny")
~/guard-lab/code/np-tests/np-matrix.sh matrix.tsv
# As a Markdown table for the CI summary
~/guard-lab/code/np-tests/np-matrix.sh --md matrix.tsv >> "$GITHUB_STEP_SUMMARY"
# A different conncheck.sh (e.g. in the CI checkout)
CC=$PWD/code/demo/conncheck.sh code/np-tests/np-matrix.sh code/np-tests/matrix.tsv
| Output / exit |
conncheck.sh |
np-matrix.sh |
| 0 |
allowed or as expected |
all lines as expected |
| 1 |
blocked or deviation |
at least one ✘ |
| 2 |
usage error, source missing |
file missing (line with exit 2 = error) |
| HTTP 403/502 |
counts as allowed |
connection was established, check L7 separately |
| Message |
Meaning |
First suspect |
Timeout |
packet dropped or DNS starved |
policy, egress DNS |
DNS-Fehler |
name does not exist (an answer came back) |
typo, namespace |
abgelehnt |
port closed, no process |
port, Service, Pod |
keine Verbindung |
TCP test failed |
like timeout |
NetworkPolicy patterns
# Default deny for everything (ingress and egress) in the namespace
spec: {podSelector: {}, policyTypes: [Ingress, Egress]}
# Allow DNS (always together with egress deny)
egress:
- to:
- namespaceSelector:
matchLabels: {kubernetes.io/metadata.name: kube-system}
podSelector:
matchLabels: {k8s-app: kube-dns}
ports: [{protocol: UDP, port: 53}, {protocol: TCP, port: 53}]
# Everything within the own namespace (ingress and egress)
ingress: [{from: [{podSelector: {}}]}]
egress: [{to: [{podSelector: {}}]}]
# Monitoring: exactly one Pod from exactly one namespace, only port metrics
ingress:
- from:
- namespaceSelector:
matchLabels: {kubernetes.io/metadata.name: monitoring}
podSelector:
matchLabels: {app: scraper}
ports: [{port: metrics}]
# Internet without cluster networks
egress:
- to: [{ipBlock: {cidr: 0.0.0.0/0, except: [10.0.0.0/8, 172.16.0.0/12,
192.168.0.0/16]}}]
ports: [{port: 443}]
# Port range (numeric only, not with a named port)
ports: [{protocol: TCP, port: 9000, endPort: 9200}]
Selector semantics
Notation under from/to |
Meaning |
Rule of thumb |
- namespaceSelector + podSelector (one
dash) |
Pods with X in namespaces with Y |
AND |
- namespaceSelector / - podSelector (two
dashes) |
namespaces with Y or Pods with X in the own NS |
OR |
- podSelector: {} |
all Pods in the own namespace |
own NS only |
- namespaceSelector: {} |
all Pods in all namespaces |
entire cluster |
ingress: [{}] |
everything from everywhere |
free pass |
policyTypes: [Egress] without egress |
nothing may leave |
deny egress |
| Rule |
Consequence |
| Policies are additive |
union of all permissions, no deny field |
| Pod selected once |
isolated for this direction (even by an “exception”) |
| Port and source |
stay coupled per rule |
| Port in the policy |
container port after DNAT (8080, not Service port 80) |
| Connection between two isolated Pods |
egress at the sender and ingress at the
receiver |
| Return packets |
need no rule (stateful) |
ipBlock for Pod IPs |
unspecified; use selectors for cluster targets |
kubernetes.io/metadata.name |
set by the API server, cannot be forged |
Cilium: CNP and CCNP
| Field |
Purpose |
Example |
endpointSelector |
protected Pods (like podSelector) |
matchLabels: {app: api} |
fromEndpoints / toEndpoints |
Pods as the other side |
k8s:io.kubernetes.pod.namespace: monitoring |
fromEntities / toEntities |
logical targets |
world, cluster, host,
kube-apiserver |
toFQDNs |
egress by domain |
matchName: example.com,
matchPattern: "*.github.com" |
toCIDR / toCIDRSet |
IP networks |
169.254.169.254/32 |
toPorts[].ports[] |
port as a string |
{port: "8080", protocol: TCP} |
toPorts[].rules.http |
L7 rules (Envoy) |
{method: GET, path: "/api/.*"} |
toPorts[].rules.dns |
DNS proxy, mandatory for toFQDNs |
matchPattern: "*" |
ingressDeny / egressDeny |
explicit denial, beats every allow |
tenant separation |
enableDefaultDeny |
false = policy closes nothing else |
guardrails in CCNP |
specs |
several rules in one object |
CCNP tenant-deny |
prefix k8s:io.cilium.k8s.namespace.labels. |
select namespace labels |
…labels.tenant |
hubble observe
hubble observe -P --verdict DROPPED --namespace shop --last 20
hubble observe -P --type policy-verdict --to-namespace shop --print-policy-names
hubble observe -P --from-pod default/client --to-port 443 -f
hubble observe -P --protocol http --to-namespace shop --http-status 403
hubble observe -P --to-fqdn "*.github.com" --last 10
hubble observe -P --namespace shop --last 200 -o json > flows.json
| Filter |
Effect |
Filter |
Effect |
--verdict DROPPED |
dropped only |
--type policy-verdict |
policy decisions only |
--from-pod ns/pod |
source |
--to-pod ns/pod |
target |
--namespace ns |
source or target |
--to-namespace ns |
target only |
--to-port 8080 |
target port (after DNAT) |
--to-ip 1.2.3.4 |
target IP |
--protocol http |
L7 HTTP flows |
--http-status 403 |
HTTP status |
--to-fqdn name |
target domain |
--label app=api |
endpoint label |
-f / --last N |
live / last N |
-o json |
export |
| Hubble output |
Meaning |
policy-verdict:none INGRESS DENIED |
no rule matches, default deny |
policy-verdict:L3-L4 … ALLOWED |
an L3/L4 rule allowed it |
Policy denied by denylist DROPPED |
deny rule (ingressDeny/egressDeny) |
http-request DROPPED |
an L7 rule denied it (client sees 403) |
ClusterNetworkPolicy (v1alpha2)
Admin tier (priority 0..1000, lowest first) ── Accept / Deny ──► done
│ Pass or no match
namespace NetworkPolicies ── Pod covered ──► Allow / default deny
│ Pod not covered by any NetworkPolicy
Baseline tier ── Accept / Deny ──► done (otherwise: allowed)
| Field |
Values |
Note |
spec.tier |
Admin, Baseline |
replaces ANP and BANP (v1alpha1) |
spec.priority |
0–1000 |
equal priority with overlap: undefined |
spec.subject |
namespaces or pods (with
namespaceSelector) |
whom the rule protects |
ingress[].from / egress[].to |
namespaces, pods, networks
(egress only) |
other side |
action |
Accept, Deny, Pass |
Pass hands off to NetworkPolicies |
protocols |
tcp: {destinationPort: {number: 9113}} |
replaces ports from v1alpha1 |
| Implementation |
Status 2026 |
| kube-network-policies |
ClusterNetworkPolicy v1alpha2 |
| Calico |
from 3.32, plus its own GlobalNetworkPolicy with
tiers |
| OVN-Kubernetes, Antrea |
ANP/BANP v1alpha1 |
| Cilium |
none, CiliumClusterwideNetworkPolicy instead |
Kyverno: policy types (policies.kyverno.io/v1)
| Type |
Short name |
Purpose |
Important fields |
ValidatingPolicy |
vpol |
check, Audit/Warn/Deny |
validations, validationActions |
MutatingPolicy |
mpol |
set defaults |
mutations (ApplyConfiguration,
JSONPatch) |
GeneratingPolicy |
gpol |
create objects |
generate, evaluation.synchronize |
ImageValidatingPolicy |
ivpol |
signatures, attestations |
matchImageReferences, attestors |
DeletingPolicy |
dpol |
scheduled deletion |
schedule, conditions |
PolicyException |
– |
time-limited exception |
policyRefs, matchConditions,
expiresAt |
ClusterPolicy (kyverno.io/v1) |
cpol |
the old JMESPath world |
deprecated in 1.19 |
Namespaced variants: nvpol, nmpol,
ngpol, nivpol, ndpol.
| Field |
Effect |
matchConstraints.resourceRules |
group, version, resource, operation |
matchConstraints.namespaceSelector |
namespace labels (for namespaces: their own labels) |
matchConstraints.objectSelector |
labels of the object itself |
matchConditions |
CEL pre-filter, separate “finegrained” webhook |
variables |
intermediate values, access via
variables.<name> |
validations[].message /
messageExpression |
fixed or computed message |
validationActions |
[Audit], [Warn], [Deny],
[Audit, Warn] |
failurePolicy |
Fail (default) or Ignore when Kyverno is
down |
webhookConfiguration.timeoutSeconds |
1–30, after that failurePolicy applies |
evaluation.background.enabled |
background scan on/off |
evaluation.generateExisting.enabled |
GeneratingPolicy for existing objects too |
autogen.podControllers.controllers |
control autogen for Deployments & co. |
autogen.validatingAdmissionPolicy.enabled |
additionally generate a native VAP |
CEL snippets
# Optional fields without errors
object.spec.?hostNetwork.orValue(false)
object.metadata.?labels[?'app'].orValue('') # keys with - or /
# All containers incl. init and ephemeral
object.spec.containers + object.spec.?initContainers.orValue([])
+ object.spec.?ephemeralContainers.orValue([])
# Collect violations, message with names
object.spec.containers.filter(c, !has(c.resources)).map(c, c.name).join(', ')
# Compare quantities (never as strings)
quantity(c.resources.limits['memory']).isGreaterThan(quantity('512Mi'))
# Image library (Kyverno)
image(i).registry() in ['index.docker.io', 'kind-registry:5000']
!image(i).containsDigest() && image(i).tag() == 'latest'
# CIDR (Kubernetes CEL)
cidr(p.ipBlock.cidr).prefixLength() >= 16
# Time (Kyverno: time.now(), not now())
timestamp(object.spec.expiresAt) < time.now() + duration('2160h')
# Live data from the cluster (type any: dyn() before exists/all)
resource.Get("v1", "configmaps", "kyverno", "allowed-registries").data
dyn(resource.List('networking.k8s.io/v1', 'networkpolicies', ns).items)
# Mutation via ApplyConfiguration
Object{metadata: Object.metadata{labels: {"guard-lab/defaults": "applied"}}}
# Generate
generator.Apply(object.metadata.name, [variables.source])
Kyverno CLI
# Apply a policy to manifests (namespace labels via -f/--values-file)
kyverno apply policy.yaml -r pods.yaml -f values.yaml --remove-color
kyverno apply policy.yaml -r ~/guard-lab/code/demo/ -f values.yaml -t
# With an exception, as a PolicyReport
kyverno apply policy.yaml -r pods.yaml --exceptions exception.yaml
kyverno apply policy.yaml -r pods.yaml -p --output-format json
# Context for resource.List/Get offline
kyverno apply policy.yaml -r ns.yaml --context-file context.yaml
# Unit tests: every kyverno-test.yaml, recursively
kyverno test .
kyverno test tests/image-hygiene --remove-color --detailed-results
kyverno test . -o junit # mixes logs into stdout, see ci/kyverno-junit.sh
# Migrate stored versions
kyverno migrate --help
Result in kyverno test |
Meaning |
Pass / Ok |
expectation met, object really checked |
Pass / Excluded |
object does not match the selection, always counts
as passed |
Fail / Want fail, got pass |
expectation missed |
RULE empty |
CEL policy, hence isValidatingPolicy: true instead of
rule: |
Chainsaw
chainsaw lint test -f e2e/image-hygiene/chainsaw-test.yaml
chainsaw test e2e/ --report-format JUNIT-TEST --report-name chainsaw
chainsaw test e2e/ --skip-delete # keep the namespace for debugging
spec:
namespaceTemplate:
metadata: {labels: {guard-lab: demo}} # so that namespaceSelector matches
steps:
- try:
- apply: {file: pod-latest.yaml, expect: [{check: {($error != null): true}}]}
- error: {file: pod-latest.yaml} # object must not exist
- assert: {file: pod-good.yaml} # object must exist like this
Troubleshooting
# NetworkPolicies: what applies to this Pod?
kubectl -n shop get netpol
kubectl -n shop describe netpol api-from-frontend | sed -n '/Spec/,$p'
kubectl get ns --show-labels
kubectl -n shop exec deploy/frontend -- nslookup api.shop.svc.cluster.local
# Cilium
cilium status
kubectl -n kube-system exec ds/cilium -c cilium-agent -- cilium-dbg status
kubectl -n kube-system exec ds/cilium -c cilium-agent -- cilium-dbg fqdn cache list
kubectl -n kube-system exec ds/cilium -c cilium-agent -- cilium-dbg identity get 4012
kubectl get cnp,ccnp -A
# Kyverno: is it running, what is it blocking?
kubectl -n kyverno get deploy,pdb
kubectl -n kyverno get endpointslice -l kubernetes.io/service-name=kyverno-svc
kubectl get validatingwebhookconfigurations -l webhook.kyverno.io/managed-by=kyverno
kubectl get vpol,mpol,gpol,ivpol,dpol
kubectl -n kyverno logs deploy/kyverno-admission-controller --since=10m | grep ERR
kubectl get polr -A; kubectl get clusterpolicyreports
kubectl get events -A --field-selector reason=PolicyViolation
kubectl run t -n shop -l app=t --image=nginxinc/nginx-unprivileged:latest \
--dry-run=server -o name
# Find generated objects
kubectl get netpol -A -l app.kubernetes.io/managed-by=kyverno \
-L generate.kyverno.io/policy-name
# Emergency exit for a dead Kyverno (then restart it; the webhooks come back)
kubectl delete validatingwebhookconfiguration,mutatingwebhookconfiguration \
-l webhook.kyverno.io/managed-by=kyverno