gRPC & High Performance APIs · 课时

使用服务网格管理 gRPC 流量

使用 Istio 等服务网格，通过边车代理、智能路由、重试和金丝雀发布来管理云中的 gRPC 流量。

使用服务网格管理 gRPC 流量 是 CoddyKit 上的免费 gRPC & High Performance APIs 课时。这是第 4 节课，共 4 节。你可以在下方免费阅读本课时的完整内容 — 然后在浏览器中使用内置代码编辑器和全天候 AI 导师进行实践。这是 gRPC & High Performance APIs 学习路径的一部分，你的进度在网页和 CoddyKit 应用中同步。 gRPC & High Performance APIs 课程共包含 4 节课。

本课时的部分内容尚未翻译，以英文显示。

Why a Service Mesh

As gRPC microservices multiply, cross-cutting concerns — routing, retries, mTLS, observability — clutter every service. A service mesh moves these into the platform layer.

The Sidecar Proxy

A mesh injects a sidecar proxy (e.g. Envoy) next to each pod. All gRPC traffic flows through it, so the proxy can apply policy without changing app code.

Data Plane vs Control Plane

Two halves:

Data plane: the proxies that move traffic
Control plane: configures the proxies (Istiod in Istio)

You write high-level rules; the control plane programs each proxy.

gRPC-Aware Load Balancing

Plain L4 load balancers pin all streams of a connection to one backend, unbalancing gRPC. The mesh proxy is L7 and balances per-request, spreading RPCs evenly.

Defining a VirtualService

An Istio VirtualService routes traffic by rules — host, headers, or weights. Here all traffic goes to the v1 subset.

spec:
  hosts: ['orders']
  http:
  - route:
    - destination: { host: orders, subset: v1 }

Canary Releases by Weight

Split traffic by percentage to roll out a new version gradually and watch its error rate before shifting more.

route:
- destination: { host: orders, subset: v1 }
  weight: 90
- destination: { host: orders, subset: v2 }
  weight: 10

Automatic Retries

The mesh can retry failed RPCs without app code. Configure attempts and per-try timeout, and limit retries to safe (idempotent) methods.

retries:
  attempts: 3
  perTryTimeout: 2s
  retryOn: 'unavailable'

Timeouts and Circuit Breaking

Set request timeouts and outlier detection so unhealthy backends are ejected automatically, protecting the system from cascading failures.

outlierDetection:
  consecutive5xxErrors: 5
  interval: 30s

Built-in mTLS

The mesh issues and rotates certificates, encrypting all service-to-service gRPC with mutual TLS automatically — no app changes required.

Observability for Free

Because all traffic passes through proxies, the mesh emits metrics, distributed traces, and access logs for every gRPC call, feeding dashboards like Kiali and Grafana.

Trade-offs

Meshes add value but also cost:

Extra latency per hop via the sidecar
Higher resource usage
Operational complexity

Adopt when service count and policy needs justify it.

Quick Check

Test your mesh knowledge.

Recap

You learned mesh traffic management:

Sidecar proxies apply policy without app changes
L7 proxies load-balance gRPC per request
VirtualService rules enable canary releases by weight
Built-in retries, timeouts, circuit breaking, and mTLS
Free observability — at the cost of latency and complexity

常见问题解答

「使用服务网格管理 gRPC 流量」课时是免费的吗？

是的 — 「使用服务网格管理 gRPC 流量」的完整文本可在网页上免费阅读。要进行交互式练习（内置代码编辑器和全天候 AI 导师）并解锁 gRPC & High Performance APIs 课程的其余内容，请升级到 CoddyKit PRO。 gRPC & High Performance APIs 课程共包含 4 节课。

「使用服务网格管理 gRPC 流量」这节课中我会学到什么？

使用 Istio 等服务网格，通过边车代理、智能路由、重试和金丝雀发布来管理云中的 gRPC 流量。你通过在浏览器中直接运行的动手代码来练习 gRPC & High Performance APIs，全天候 AI 导师会在你学习这节课的过程中回答你的问题。

学习 gRPC & High Performance APIs 需要有经验吗？

无需任何先前经验。CoddyKit 上的 gRPC & High Performance APIs 课程适合初学者到高级学习者，你可以从这里开始或从头开始，按照自己的节奏学习。这是第 4 节课，共 4 节。

「使用服务网格管理 gRPC 流量」课时需要多长时间？

大多数 CoddyKit 课程大约需要 5–10 分钟。每节课都很精短且互动，所以你能稳步进步，并在网页和应用中从离开的地方继续。

我能在这节 gRPC & High Performance APIs 课中编写并运行代码吗？

能。每节 gRPC & High Performance APIs 课都包含内置代码编辑器，你可以在浏览器中直接编写并运行真实代码，并获得即时 AI 反馈 — 无需本地设置。

此课程中的所有课时

← 返回 gRPC & High Performance APIs