Files
schemas/k8s/deploy.yaml
T
argoyle 3709a90124
schemas / vulnerabilities (push) Successful in 1m2s
schemas / check (push) Successful in 1m9s
Unbound Release / Check Preconditions (push) Successful in 2s
Unbound Release / Create Release (push) Skipped
schemas / check-release (push) Successful in 1m23s
pre-commit / pre-commit (push) Successful in 1m47s
Unbound Release / Generate Changelog and Handle PR (push) Successful in 10s
Unbound Release / Create Tag (push) Successful in 6s
Release / release (push) Successful in 2m24s
schemas / build (push) Successful in 2m10s
schemas / deploy-prod-frostmoln (push) Successful in 1m2s
feat: deploy to Frostmoln production (#1019)
Moves the production deploy from AWS to the Frostmoln production cluster (namespace `schemas`).

- `deploy-prod` (AWS) has failed since CI left AWS (no `k8s.unbound.se` context on the runners). It's replaced by `deploy-prod-frostmoln`, as in unbound-site/lingonpaj-site: `deploy prod-frostmoln -n schemas` with the org's `FROSTMOLN_PRODUCTION_DEPLOY_KEY`, `FROSTMOLN_PRODUCTION_CLUSTER_ID` and `BUILDTOOLS_FROSTMOLN_PRODUCTION`.
- Restricted Pod Security: the image runs as `node` (uid 1000) with a matching securityContext. The service only writes under `/tmp` (cosmo composition).
- `secrets.yaml` (AWS ClusterSecretStore, RDS + CloudAMQP) → `secrets-prod-frostmoln.yaml`: the namespace's `frostmoln` SecretStore and the Frostmoln Secret `schemas` (`POSTGRES_URL`, `AMQP_URL`, `API_KEY`), written by infra's `seed-schemas.sh`. The file had no target suffix, so it applied to every target.
- `config-prod`/`pdb-prod` → `-prod-frostmoln`; the ALB ingress → a Traefik one for `/query` with a Let's Encrypt certificate.

Checked locally (amd64): runs as uid 1000 with all capabilities dropped, `wgc` works, `/tmp` writable; it fails only on the (dummy) database, as expected.

**Merge after** infra#1878 is applied and `seed-schemas.sh` + `migrate-schemas-db.sh` have run. `schemas.unbound.se` keeps pointing at AWS until the cut-over.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

https://claude.ai/code/session_01Ma8eHhW59hLAJuLzBnLvUu
Reviewed-on: #1019
2026-10-15 21:15:30 +00:00

104 lines
2.4 KiB
YAML

apiVersion: v1
kind: ServiceAccount
metadata:
name: schemas
---
apiVersion: apps/v1
kind: Deployment
metadata:
labels:
app.kubernetes.io/name: schemas
name: schemas
annotations:
kubernetes.io/change-cause: "${TIMESTAMP} Deployed commit id: ${COMMIT}"
spec:
replicas: 1
selector:
matchLabels:
app.kubernetes.io/name: schemas
strategy:
rollingUpdate:
maxSurge: 1
maxUnavailable: 1
type: RollingUpdate
template:
metadata:
labels:
app.kubernetes.io/name: schemas
spec:
securityContext:
runAsNonRoot: true
runAsUser: 1000
seccompProfile:
type: RuntimeDefault
affinity:
podAntiAffinity:
preferredDuringSchedulingIgnoredDuringExecution:
- weight: 100
podAffinityTerm:
labelSelector:
matchExpressions:
- key: "app.kubernetes.io/name"
operator: In
values:
- schemas
topologyKey: kubernetes.io/hostname
containers:
- name: schemas
securityContext:
allowPrivilegeEscalation: false
capabilities:
drop: ["ALL"]
resources:
requests:
cpu: "100m"
memory: "128Mi"
limits:
memory: "768Mi"
livenessProbe:
httpGet:
path: /health/live
port: 8080
initialDelaySeconds: 10
periodSeconds: 10
timeoutSeconds: 5
failureThreshold: 3
readinessProbe:
httpGet:
path: /health/ready
port: 8080
initialDelaySeconds: 5
periodSeconds: 5
timeoutSeconds: 5
failureThreshold: 3
imagePullPolicy: IfNotPresent
image: ${IMAGE}
ports:
- name: api
containerPort: 8080
env:
- name: OTEL_EXPORTER_OTLP_ENDPOINT
value: http://k8s-monitoring-alloy.monitoring.svc.cluster.local:4318
envFrom:
- configMapRef:
name: schemas
- secretRef:
name: schemas
restartPolicy: Always
serviceAccountName: schemas
---
apiVersion: v1
kind: Service
metadata:
name: schemas
spec:
ports:
- port: 80
name: api
protocol: TCP
targetPort: 8080
selector:
app.kubernetes.io/name: schemas
type: NodePort