Service Mesh
Architecture
┌─────────────────────────────┐
│ Control Plane │ (Istio: istiod / Linkerd: control plane)
└──────────────┬──────────────┘
│ config push
┌──────────┼──────────┐
▼ ▼ ▼
┌────────┐ ┌────────┐ ┌────────┐
│ Sidecar│ │ Sidecar│ │ Sidecar│ (Envoy / linkerd-proxy)
│ Proxy │ │ Proxy │ │ Proxy │
├────────┤ ├────────┤ ├────────┤
│ App A │ │ App B │ │ App C │
└────────┘ └────────┘ └────────┘
Istio Traffic Management
# VirtualService: route rules
apiVersion: networking.istio.io/v1beta1
kind: VirtualService
metadata:
name: order-service
spec:
hosts: [order-service]
http:
- match:
- headers:
x-canary: { exact: "true" }
route:
- destination: { host: order-service, subset: canary }
- route:
- destination: { host: order-service, subset: stable }
weight: 90
- destination: { host: order-service, subset: canary }
weight: 10
# DestinationRule: subsets + circuit breaker
apiVersion: networking.istio.io/v1beta1
kind: DestinationRule
metadata:
name: order-service
spec:
host: order-service
trafficPolicy:
connectionPool:
tcp: { maxConnections: 100 }
http: { h2UpgradePolicy: DEFAULT, http1MaxPendingRequests: 100 }
outlierDetection:
consecutive5xxErrors: 5
interval: 30s
baseEjectionTime: 30s
subsets:
- name: stable
labels: { version: v1 }
- name: canary
labels: { version: v2 }
Mutual TLS
# PeerAuthentication: enforce mTLS
apiVersion: security.istio.io/v1beta1
kind: PeerAuthentication
metadata:
name: default
namespace: production
spec:
mtls:
mode: STRICT # All traffic must be mTLS
Linkerd (simpler alternative)
# Install
linkerd install | kubectl apply -f -
# Inject sidecar into deployment
kubectl get deploy order-service -o yaml | linkerd inject - | kubectl apply -f -
# Traffic split
apiVersion: split.smi-spec.io/v1alpha2
kind: TrafficSplit
metadata:
name: order-canary
spec:
service: order-service
backends:
- service: order-service-stable
weight: 900
- service: order-service-canary
weight: 100
When to Use a Service Mesh
| Use When | Don't Use When |
|---|---|
| 10+ microservices | Monolith or few services |
| Need mTLS everywhere | TLS at ingress is sufficient |
| Complex traffic routing | Simple load balancing works |
| Multi-team ownership | Single team manages all services |
Anti-Patterns
| Anti-Pattern | Fix |
|---|---|
| Service mesh for 2-3 services | Overkill; use in-app libraries |
| No resource limits on sidecars | Configure proxy CPU/memory limits |
| mTLS in PERMISSIVE mode in prod | Use STRICT mode in production |
| No observability dashboards | Deploy Kiali, Grafana, Jaeger with mesh |
| Ignoring sidecar latency | Benchmark; typically adds <1ms per hop |
Production Checklist
- mTLS in STRICT mode
- Circuit breaker policies on critical services
- Traffic splitting for canary deployments
- Observability: Kiali dashboard, distributed tracing
- Resource limits on sidecar proxies
- Gradual rollout (inject namespace by namespace)