Production Launch Checklist
Run through these checks before cutting over to production traffic. Items marked with "auto" are verified by test-deployment.sh.
1. TLS Certificates
Status: Verify certificate is issued and not expiring soon.
kubectl get certificate -n platform -o wide
kubectl describe certificate platform-tls -n platform
kubectl get certificaterequest -n platform
| Check |
Command |
Expected |
| Certificate Ready |
kubectl get certificate -n platform |
READY=True |
| Expiry |
kubectl get certificate -n platform -o jsonpath='{.status.notAfter}' |
> 30 days |
| Auto-renewal |
Check cert-manager logs |
cert-manager/controller logs show renewal check |
| Staging vs Prod |
Verify ClusterIssuer |
letsencrypt-prod, not letsencrypt-staging |
2. Database Backups
Status: Verify backup infrastructure and perform a test restore.
# Check MinIO connectivity
curl -I http://minio.storage:9000
# Verify backup provider is registered
curl -H "Authorization: Bearer $TOKEN" https://$DOMAIN/api/settings/storage
# Trigger a test backup
curl -X POST -H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/json" \
-d '{"dbName":"platform","environment":"production"}' \
https://$DOMAIN/api/projects/{projectId}/databases/backup
| Check |
Details |
| Storage provider configured |
Default MinIO or external S3 |
| Backup retention policy |
Defined in backup schedule |
| Test restore performed |
Restore from backup to isolated DB |
| Backup monitoring |
Alert on backup failure |
3. MinIO / S3 Storage
Status: Object storage is accessible and has sufficient free space.
kubectl get pods -n storage
kubectl get pvc -n storage
| Check |
Expected |
| MinIO pod Running |
kubectl get pods -n storage → Running |
| PVC bound |
kubectl get pvc -n storage → Bound |
| Free space |
kubectl exec -n storage pod/minio-xxx -- df -h /data |
| External S3 fallback |
If MinIO fails, external S3 configured |
| Default buckets exist |
platform-backups, platform-logs |
4. SMTP Configuration
Status: Email delivery is configured and tested.
# Test SMTP configuration via API
curl -X POST -H "Authorization: Bearer $TOKEN" \
https://$DOMAIN/api/settings/smtp/{id}/test \
-H "Content-Type: application/json" \
-d '{"testTo":"[email protected]"}'
| Check |
Details |
| SMTP provider configured |
Custom SMTP, SES, SendGrid, or Mailgun |
| Test email sent |
Verify inbox receives notification |
| From address verified |
SPF/DKIM records configured for domain |
| SMTP credentials valid |
Not expired or revoked |
| Fallback provider |
(optional) Secondary SMTP |
5. Monitoring Dashboards
Status: All dashboards load and display data.
# Verify Grafana is accessible
curl -sI https://$DOMAIN/grafana/
# Check Prometheus targets
curl -s https://$DOMAIN/api/health
| Check |
Details |
| Grafana accessible |
HTTPS 200/302 |
| Prometheus targets healthy |
kubectl get pods -n monitoring |
| Loki receiving logs |
Grafana Explore → Loki → run Log query |
| Pre-built dashboards loaded |
Platform Overview, API Performance, Service Health |
| OIDC SSO works |
Log in via Cairn credentials |
6. RBAC & Roles
Status: Role presets and custom roles are configured, users assigned.
# Check existing roles
curl -H "Authorization: Bearer $TOKEN" https://$DOMAIN/api/roles
# Verify user permissions
curl -H "Authorization: Bearer $TOKEN" https://$DOMAIN/api/users/{id}/permissions
| Check |
Details |
| Admin user exists |
[email protected] or equivalent |
| DevOps users configured |
At least 2 for redundancy |
| Custom roles defined |
If needed per team structure |
| Viewer accounts for stakeholders |
Read-only access |
No orphaned roleId references |
Custom roles that were deleted |
7. Audit Logging
Status: All sensitive operations are logged.
# View recent audit logs
curl -H "Authorization: Bearer $TOKEN" https://$DOMAIN/api/audit-logs
| Check |
Details |
| Audit logs accessible |
200 response |
| Secret operations logged |
secrets.create, secrets.reveal, secrets.delete |
| User management logged |
user.created, user.deleted, user.role-updated |
| Deployment operations logged |
deployment.created, deployment.rolled_back |
| Log retention policy |
Defined (default: indefinite in PostgreSQL) |
8. Resource Limits
Status: All deployments have resource requests/limits configured.
# Check resource limits across namespaces
kubectl get pods -A -o json | jq '.items[] | {name: .metadata.name, ns: .metadata.namespace, limits: .spec.containers[0].resources.limits, requests: .spec.containers[0].resources.requests}'
| Check |
Details |
| API has resource limits |
CPU/Memory requests and limits |
| Portal has resource limits |
CPU/Memory requests and limits |
| PostgreSQL has limits |
Especially memory limit |
| MongoDB has limits |
Especially memory limit |
| Preview namespaces limited |
ResourceQuota enforced for preview namespace |
Recommended Minimum Limits
| Service |
CPU Request |
CPU Limit |
Memory Request |
Memory Limit |
| Cairn API |
100m |
500m |
256Mi |
512Mi |
| Cairn portal |
50m |
200m |
128Mi |
256Mi |
| PostgreSQL |
500m |
2000m |
1Gi |
4Gi |
| MongoDB |
500m |
2000m |
1Gi |
4Gi |
| Redis |
100m |
500m |
256Mi |
512Mi |
| MinIO |
200m |
1000m |
512Mi |
2Gi |
| Preview (per env) |
50m |
200m |
128Mi |
256Mi |
9. DNS Records
Status: All DNS records are configured and resolving.
# Verify DNS resolution
nslookup $DOMAIN
dig $DOMAIN +short
# Check all subdomains
for sub in "" "api" "grafana" "argocd" "portainer" "minio"; do
echo "$sub.$DOMAIN → $(dig +short $sub.$DOMAIN)"
done
| Record |
Type |
Value |
{DOMAIN} |
A |
Server public IP |
*.{DOMAIN} |
A |
Server public IP (wildcard) |
Additional Checks
| Check |
Details |
| Wildcard DNS configured |
Required for preview URLs (*.preview.{DOMAIN}) |
| Reverse DNS set |
PTR record if required |
| TTL reasonable |
300–600 seconds (before launch), higher after |
| HTTPS reachable |
curl -sI https://$DOMAIN returns 200 |
| No mixed content |
All resources served over HTTPS |