Skip to content

deployment.md

Latest commit

 

History

History
165 lines (132 loc) · 3.95 KB

deployment.md

File metadata and controls

165 lines (132 loc) · 3.95 KB

MiniRedis - Deployment and DevOps Plan

1. Deployment Environments

Local (Developer)

  • Tooling: Docker Compose
  • Purpose: feature development and integration validation
  • Characteristics:
    • direct host port mapping
    • local .env-driven config
    • fast iteration

Staging

  • Tooling: Kubernetes namespace (miniredis-staging)
  • Purpose: pre-production integration and release validation
  • Characteristics:
    • production-like routing and policies
    • synthetic load and smoke tests
    • rollback rehearsal

Production

  • Tooling: Kubernetes namespace (miniredis-prod)
  • Purpose: live workloads
  • Characteristics:
    • managed TLS, secret management, observability stack
    • autoscaling and resilience controls
    • progressive delivery

2. Container and Build Strategy

  • One service per image.
  • Multi-stage Docker builds.
  • Immutable tags:
    • service:<git-sha>
    • optional semver tags for releases.
  • Registry push from CI after tests pass.
  • Runtime images kept minimal for attack-surface reduction.

3. Kubernetes Design

3.1 Core Workloads

  • Deployment:
    • frontend
    • api-gateway
    • backend
    • node-manager
    • monitoring-service
    • auth-service
  • StatefulSet:
    • postgres-main (if self-hosted)
    • postgres-auth (if self-hosted)

3.2 Core Objects

  • Service (ClusterIP) for internal communication.
  • Ingress for external access.
  • ConfigMap for non-sensitive config.
  • Secret for credentials/tokens/keys.
  • PVC for stateful storage.

4. Port and Service Mapping for K8s

Logical Service App Port K8s Service Port Public Exposure
frontend 5173 80/443 via ingress public
api-gateway 8080 8080 public/internal
backend 5500 5500 internal
node-manager 7000 7000 internal
monitoring-service 9000 9000 internal via gateway or restricted public
auth-service 8000 8000 internal via gateway
postgres-main 5432 5432 internal only
postgres-auth 5432 5432 internal only

Tenant node port pool:

  • controlled by env (REDIS_PORT_START, REDIS_PORT_END)
  • verify no conflict with node/host networking strategy.

5. Reliability Standards

Each service should have:

  • readiness probe,
  • liveness probe,
  • CPU/memory requests and limits,
  • rolling update strategy.

Critical services should also have:

  • PodDisruptionBudget,
  • anti-affinity rules,
  • HPA (api-gateway, backend, monitoring).

6. Security Controls

  • Namespace segmentation by environment.
  • NetworkPolicy default deny + explicit allow.
  • TLS managed with cert-manager.
  • Secrets from secret manager (or sealed secrets).
  • Non-root container execution where possible.
  • CI image and dependency scanning.

7. Observability Stack

  • Metrics: Prometheus
  • Dashboards: Grafana
  • Logs: Loki
  • Traces: OpenTelemetry collector (planned)

Minimum dashboards:

  • gateway latency/error rate
  • backend request/error trends
  • node-manager active node count + memory
  • monitoring freshness and scrape health
  • database health and connection saturation

8. CI/CD Pipeline (Target)

CI

  1. Lint and static checks.
  2. Unit tests.
  3. Integration tests (compose/k8s ephemeral).
  4. Build and scan images.
  5. Push versioned images.

CD (GitOps)

  1. Update Helm values/image tags.
  2. Argo CD sync to staging.
  3. Smoke test gate.
  4. Promote to production.
  5. Monitor rollback triggers.

9. Rollout and Rollback

  • Default: rolling update.
  • Sensitive services: canary rollout.
  • Auto rollback triggers:
    • readiness failures,
    • elevated 5xx rates,
    • health-check failure budget breach.

10. Backup/Recovery Plan

  • Daily full backups for DB.
  • Periodic restore test in staging.
  • RPO/RTO documented for production.
  • Versioned migration policy for schema changes.

11. Deployment Milestones

  1. Helm charts for all current services.
  2. Staging namespace with ingress + TLS.
  3. Observability baseline in cluster.
  4. GitOps promotion workflow.
  5. Production hardening and runbooks.