kubernetes-patterns
Pods, deployments, services, ingress, RBAC, autoscaling, and production cluster best practices
Informations de source
- Dépôt
- cosmicstack-labs/mercury-agent-skills
- Dernière activité de la source
- 10 mai 2026 à 06:29
- Langue détectée de SKILL.md
- anglais
- Étoiles
- 476
- Forks
- 63
Options d'installation
Le prompt qui vérifie d'abord la source est sélectionné par défaut. Vous pouvez passer à une commande directe ou télécharger une copie locale.
Vérifiez les fichiers source
Lisez SKILL.md et les fichiers associés affichés par SkillsMP avant de décider de l'installer.
Affichage de SKILL.md
SKILL.md
Instructions source · Aperçu en lecture seule- name
- kubernetes-patterns
- description
- Pods, deployments, services, ingress, RBAC, autoscaling, and production cluster best practices
- metadata
- {"author":"cosmicstack-labs","version":"1.0.0","category":"devops","tags":["kubernetes","k8s","devops","containers","orchestration"]}
# Kubernetes Patterns
Deploy and manage production Kubernetes clusters.
## Core Objects
### Workloads
| Object | Use Case | Scaling |
|--------|----------|---------|
| Deployment | Stateless apps | Replicas, HPA |
| StatefulSet | Stateful apps (DBs) | Stable network IDs |
| DaemonSet | Per-node agents (logging, monitoring) | Node count |
| Job/CronJob | Batch tasks, scheduled jobs | Completion |
### Networking
- **Service**: Stable endpoint for pods (ClusterIP, NodePort, LoadBalancer)
- **Ingress**: HTTP routing, TLS termination, path-based routing
- **Network Policies**: Pod-level firewall rules
## Production Best Practices
### Resource Management
```yaml
resources:
requests:
cpu: 100m
memory: 128Mi
limits:
cpu: 500m
memory: 256Mi
```
- Always set requests AND limits
- Use LimitRange for namespace defaults
- Use ResourceQuota for namespace caps
### Pod Anti-Affinity
Spread pods across nodes for HA.
### Readiness & Liveness Probes
- Readiness: traffic starts flowing
- Liveness: pod gets restarted
- Startup: for slow-starting containers
## RBAC
- Least-privilege service accounts per app
- Namespace-scoped roles (not cluster-wide)
- Regularly audit permissions
- Use groups, not individual users
## Autoscaling
- HPA: scale by CPU/memory or custom metrics
- VPA: adjust resource requests automatically
- Cluster Autoscaler: add/remove nodes
- KEDA: event-driven scaling (SQS queue depth, etc.)
Voir sur GitHub