> For the complete documentation index, see [llms.txt](https://atomoh.gitbook.io/kubernetes/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://atomoh.gitbook.io/kubernetes/es/fu-wu-wang-ge/istio/traffic-management/06-load-balancing.md).
# 负载均衡
Istio 通过 Envoy 提供多种负载均衡算法，以高效分配流量。
## 目录
1. [为什么需要负载均衡？](#why-load-balancing)
2. [负载均衡概述](#load-balancing-overview)
3. [负载均衡算法](#load-balancing-algorithms)
4. [一致性哈希详解](#consistent-hash-details)
5. [基于位置的负载均衡](#locality-based-load-balancing)
6. [连接池设置](#connection-pool-settings)
7. [实践示例](#practical-examples)
8. [算法选择指南](#algorithm-selection-guide)
9. [最佳实践](#best-practices)
10. [故障排除](#troubleshooting)
## 为什么需要负载均衡？
### 高效的资源利用
负载均衡将流量分配到多个实例，从而提高整体系统吞吐量和稳定性。
```mermaid
flowchart TB
subgraph Without["Without Load Balancing"]
C1[All Requests] -->|Overload| S1[Service 1
100% Load]
S2[Service 2
0% Load]
S3[Service 3
0% Load]
end
subgraph With["With Load Balancing"]
C2[Distributed Requests] --> LB[Load Balancer]
LB -->|33%| S4[Service 1
33% Load]
LB -->|33%| S5[Service 2
33% Load]
LB -->|34%| S6[Service 3
34% Load]
end
%% Style definitions
classDef overload fill:#FF6B6B,stroke:#333,stroke-width:1px,color:white;
classDef idle fill:#95A5A6,stroke:#333,stroke-width:1px,color:white;
classDef balanced fill:#00C7B7,stroke:#333,stroke-width:1px,color:white;
classDef lb fill:#326CE5,stroke:#333,stroke-width:1px,color:white;
classDef client fill:#f9f9f9,stroke:#333,stroke-width:1px,color:black;
%% Class applications
class C1,C2 client;
class S1 overload;
class S2,S3 idle;
class LB lb;
class S4,S5,S6 balanced;
```
### 主要优势
| 问题 | 没有负载均衡 | 使用负载均衡 |
| --------- | -------------- | ------- |
| **可用性** | 单点故障（SPOF） | 故障时自动切换 |
| **性能** | 特定实例过载 | 负载均匀分配 |
| **可扩展性** | 难以水平扩展 | 易于横向扩容 |
| **响应时间** | 不一致（0-1000ms+） | 响应时间一致 |
| **资源利用率** | 效率低（仅部分使用） | 高效使用资源 |
## 负载均衡概述
```mermaid
flowchart TB
Client[Client Request]
subgraph LB["Load Balancer"]
Algorithm[Load Balancing
Algorithm]
end
subgraph Pods["Pods"]
Pod1[Pod 1
Load: 30%]
Pod2[Pod 2
Load: 50%]
Pod3[Pod 3
Load: 20%]
end
Client --> Algorithm
Algorithm -->|Round Robin| Pod1
Algorithm -->|Least Request| Pod3
Algorithm -->|Random| Pod2
%% Style definitions
classDef client fill:#f9f9f9,stroke:#333,stroke-width:1px,color:black;
classDef lb fill:#326CE5,stroke:#333,stroke-width:1px,color:white;
classDef pod fill:#00C7B7,stroke:#333,stroke-width:1px,color:white;
%% Class applications
class Client client;
class Algorithm lb;
class Pod1,Pod2,Pod3 pod;
```
## 负载均衡算法
Istio 提供以下负载均衡算法。
### 算法比较
| 算法 | 说明 | 使用场景 | 优点 | 缺点 |
| -------------------- | --------- | ------------- | ----- | --------- |
| **ROUND\_ROBIN** | 顺序分配（默认） | 无状态服务 | 简单、公平 | 可能发生负载不均衡 |
| **LEAST\_REQUEST** | 最少活跃请求数 | 高性能 API、DB 连接 | 负载均衡 | 略有开销 |
| **RANDOM** | 随机分配 | 高流量 | 简单、快速 | 短期内可能不均衡 |
| **PASSTHROUGH** | 原始目标地址 | TCP 代理、SNI 路由 | 灵活性 | 控制有限 |
| **CONSISTENT\_HASH** | 基于哈希的粘性路由 | 会话持久化、缓存 | 粘性会话 | 可能不均衡 |
### 1. ROUND\_ROBIN（默认）
将请求按顺序分配到各个端点。
```mermaid
sequenceDiagram
autonumber
participant Client
participant LB as Load Balancer
(ROUND_ROBIN)
participant Pod1 as Pod 1
participant Pod2 as Pod 2
participant Pod3 as Pod 3
Client->>LB: Request 1
LB->>Pod1: Route
Pod1-->>Client: Response
Client->>LB: Request 2
LB->>Pod2: Route
Pod2-->>Client: Response
Client->>LB: Request 3
LB->>Pod3: Route
Pod3-->>Client: Response
Client->>LB: Request 4
LB->>Pod1: Route (cycle)
Pod1-->>Client: Response
```
**配置示例：**
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: reviews-round-robin
spec:
host: reviews
trafficPolicy:
loadBalancer:
simple: ROUND_ROBIN
```
**使用场景：**
* 无状态 REST API
* 性能相同的 Pod
* 默认设置已足够时
**优点：**
* 实现简单且可预测
* 分配公平
**缺点：**
* 不考虑每个 Pod 的负载差异
* 长时间请求可能导致不均衡
### 2. LEAST\_REQUEST
路由到活跃请求数最少的端点。
```mermaid
flowchart LR
Client[Client
New Request]
LB[Load Balancer
LEAST_REQUEST]
subgraph Pods[Pod Status]
Pod1[Pod 1
Active Requests: 5]
Pod2[Pod 2
Active Requests: 2 ✅]
Pod3[Pod 3
Active Requests: 8]
end
Client --> LB
LB -.->|Check| Pod1
LB -.->|Check| Pod2
LB -.->|Check| Pod3
LB -->|Route to
Least Loaded| Pod2
%% Style definitions
classDef client fill:#f9f9f9,stroke:#333,stroke-width:1px,color:black;
classDef lb fill:#326CE5,stroke:#333,stroke-width:1px,color:white;
classDef pod fill:#00C7B7,stroke:#333,stroke-width:1px,color:white;
classDef selected fill:#F8B52A,stroke:#333,stroke-width:2px,color:black;
%% Class applications
class Client client;
class LB lb;
class Pod1,Pod3 pod;
class Pod2 selected;
```
**配置示例：**
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: api-least-request
spec:
host: api-service
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST
warmupDurationSecs: 60 # 60 second warmup (optional)
```
**高级配置：**
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: api-least-request-advanced
spec:
host: api-service
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST
warmupDurationSecs: 120 # New pod warmup
connectionPool:
http:
http2MaxRequests: 100
maxRequestsPerConnection: 10
```
**使用场景：**
* 响应时间可变的 API
* 数据库连接池
* 处理负载较重的服务
* 实时负载均衡很重要时
**优点：**
* 适应实时负载
* 提高响应时间一致性
* 吸收每个 Pod 的性能差异
**缺点：**
* 略有开销（跟踪活跃请求）
### 3. RANDOM
随机选择一个端点。
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: reviews-random
spec:
host: reviews
trafficPolicy:
loadBalancer:
simple: RANDOM
```
**使用场景：**
* 高流量（统计上均匀）
* 需要简单快速选择时
* Pod 性能相同时
**优点：**
* 选择速度非常快
* 实现简单
* 大规模下统计上均匀
**缺点：**
* 短期内可能不均衡
* 不可预测
### 4. PASSTHROUGH
直接连接到客户端指定的原始目标地址。
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: tcp-passthrough
spec:
host: "*.external-service.com"
trafficPolicy:
loadBalancer:
simple: PASSTHROUGH
```
**使用场景：**
* TCP 代理
* 基于 SNI 的路由
* 直接连接外部服务
* TLS PASSTHROUGH 模式
**优点：**
* 保留原始目标地址
* 路由灵活
**缺点：**
* 负载均衡控制有限
### 5. LEAST\_CONN（已弃用 -> LEAST\_REQUEST）
**注意**：`LEAST_CONN` 已**弃用**，并由 `LEAST_REQUEST` 替代。
**迁移：**
```yaml
# Old version (deprecated)
trafficPolicy:
loadBalancer:
simple: LEAST_CONN
# New version
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST
```
## 一致性哈希详解
一致性哈希根据特定属性路由到同一端点，以确保会话持久化。
### 一致性哈希工作原理
```mermaid
flowchart TD
subgraph Requests[Requests]
R1[User A
Cookie: abc123]
R2[User B
Cookie: def456]
R3[User A
Cookie: abc123]
end
subgraph Hash[Hash Calculation]
H1[Hash abc123
-> 0x1A2B]
H2[Hash def456
-> 0x7C8D]
H3[Hash abc123
-> 0x1A2B]
end
subgraph Ring[Consistent Hash Ring]
P1[Pod 1
0x0000-0x4FFF]
P2[Pod 2
0x5000-0x9FFF]
P3[Pod 3
0xA000-0xFFFF]
end
R1 -->|1| H1
H1 -->|2| P1
R2 -->|1| H2
H2 -->|2| P2
R3 -->|1| H3
H3 -->|2| P1
%% Style definitions
classDef request fill:#f9f9f9,stroke:#333,stroke-width:1px,color:black;
classDef hash fill:#326CE5,stroke:#333,stroke-width:1px,color:white;
classDef pod fill:#00C7B7,stroke:#333,stroke-width:1px,color:white;
%% Class applications
class R1,R2,R3 request;
class H1,H2,H3 hash;
class P1,P2,P3 pod;
```
### 1. 基于 HTTP Header
根据特定 HTTP Header 值计算哈希。
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: user-hash
spec:
host: api-service
trafficPolicy:
loadBalancer:
consistentHash:
httpHeaderName: "x-user-id"
```
**使用场景：**
* 每用户会话持久化
* 基于 API key 的路由
* 每租户隔离
### 2. 基于 HTTP Cookie
根据 Cookie 值计算哈希；如果 Cookie 缺失，则自动生成 Cookie。
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: cookie-hash
spec:
host: web-service
trafficPolicy:
loadBalancer:
consistentHash:
httpCookie:
name: "user-session"
ttl: 3600s # 1 hour TTL
```
**使用场景：**
* Web 应用程序会话持久化
* 购物车持久化
* 用户体验一致性
### 3. 基于源 IP
根据客户端源 IP 地址计算哈希。
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: source-ip-hash
spec:
host: api-service
trafficPolicy:
loadBalancer:
consistentHash:
useSourceIp: true
```
**使用场景：**
* 基于 IP 的会话持久化
* 限流（每个 IP）
* 区域缓存
**注意事项：**
* NAT 后的客户端可能被路由到同一 Pod
* 使用代理时，请验证实际客户端 IP
### 4. 基于 HTTP 查询参数
根据查询参数值计算哈希。
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: query-param-hash
spec:
host: api-service
trafficPolicy:
loadBalancer:
consistentHash:
httpQueryParameterName: "user_id"
```
**使用场景：**
* RESTful API 中基于资源 ID 的路由
* 缓存友好的路由
* 分片策略
### 5. 最小哈希环大小设置
设置最小一致性哈希环大小，以尽量减少重新分配。
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: hash-with-ring-size
spec:
host: cache-service
trafficPolicy:
loadBalancer:
consistentHash:
httpHeaderName: "x-cache-key"
minimumRingSize: 1024 # Default: 1024
```
**说明：**
* 更大的哈希环大小可提供更均匀的分配
* 减少添加/移除 Pod 时被重新分配的 key 比例
* 内存使用量略有增加
**推荐值：**
* 小规模（< 10 个 Pod）：1024（默认）
* 中等规模（10-50 个 Pod）：2048
* 大规模（50+ 个 Pod）：4096
### 一致性哈希组合示例
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: advanced-consistent-hash
spec:
host: api-service
trafficPolicy:
loadBalancer:
consistentHash:
httpCookie:
name: "session-id"
path: "/api"
ttl: 7200s # 2 hours
minimumRingSize: 2048
connectionPool:
http:
maxRequestsPerConnection: 100
idleTimeout: 300s
```
### 一致性哈希注意事项
#### 1. 不均衡风险
```mermaid
flowchart LR
subgraph Problem[Problem Situation]
Users[1000 Users]
P1[Pod 1
800 users ⚠️]
P2[Pod 2
150 users]
P3[Pod 3
50 users]
end
Users -->|80%| P1
Users -->|15%| P2
Users -->|5%| P3
%% Style definitions
classDef user fill:#f9f9f9,stroke:#333,stroke-width:1px,color:black;
classDef overload fill:#FF6B6B,stroke:#333,stroke-width:1px,color:white;
classDef normal fill:#00C7B7,stroke:#333,stroke-width:1px,color:white;
%% Class applications
class Users user;
class P1 overload;
class P2,P3 normal;
```
**原因**：流量集中于特定哈希值
**解决方案：**
* 增加 `minimumRingSize`
* 使用多个哈希 key 组合
* 考虑 Bounded Load 算法（Envoy 不支持）
#### 2. 添加/移除 Pod 时的重新分配
```yaml
# On pod scale out
# - Before: Pod 1, Pod 2, Pod 3
# - After: Pod 1, Pod 2, Pod 3, Pod 4
# - Result: ~25% of sessions redistributed to different pods
```
**缓解策略：**
* 使用优雅关闭
* 使用外部会话存储（Redis、Memcached）
* 逐步扩缩容
## 基于位置的负载均衡
基于位置的负载均衡优先选择地理位置更近的端点。
### 基本位置配置
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: locality-lb
spec:
host: reviews
trafficPolicy:
loadBalancer:
localityLbSetting:
enabled: true
```
### 位置分配比例
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: locality-distribute
spec:
host: reviews
trafficPolicy:
loadBalancer:
localityLbSetting:
enabled: true
distribute:
- from: us-west/zone-1/*
to:
"us-west/zone-1/*": 80 # 80% to same zone
"us-west/zone-2/*": 20 # 20% to other zone
```
### 位置故障转移
当一个区域发生故障时，自动切换到另一个区域。
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: locality-failover
spec:
host: api-service
trafficPolicy:
loadBalancer:
localityLbSetting:
enabled: true
failover:
- from: us-west
to: us-east
```
### 多区域示例
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: global-service-locality
spec:
host: global-api
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST
localityLbSetting:
enabled: true
distribute:
# US West clients
- from: us-west/*
to:
"us-west/*": 90 # 90% local
"us-east/*": 10 # 10% remote (DR)
# US East clients
- from: us-east/*
to:
"us-east/*": 90
"us-west/*": 10
failover:
- from: us-west
to: us-east
- from: us-east
to: us-west
outlierDetection:
consecutiveErrors: 5
interval: 30s
baseEjectionTime: 30s
```
**使用场景：**
* 多区域部署
* 最小化延迟
* 跨区域灾难恢复
* 成本优化（同一 AZ 通信）
## 连接池设置
将连接池与负载均衡一同配置，以优化性能。
### HTTP/1.1 连接池
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: http1-connection-pool
spec:
host: api-service
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST
connectionPool:
tcp:
maxConnections: 100 # Maximum connections
connectTimeout: 3s # Connection timeout
http:
http1MaxPendingRequests: 50 # Pending request count
maxRequestsPerConnection: 100 # Max requests per connection
idleTimeout: 300s # Idle connection timeout
```
### HTTP/2 连接池
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: http2-connection-pool
spec:
host: grpc-service
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST
connectionPool:
tcp:
maxConnections: 50
http:
http2MaxRequests: 1000 # HTTP/2 concurrent requests
maxRequestsPerConnection: 0 # Unlimited (HTTP/2 multiplexing)
h2UpgradePolicy: UPGRADE # Allow HTTP/2 upgrade
```
## 实践示例
### 示例 1：高性能 API 服务
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: api-high-performance
namespace: production
spec:
host: api-service
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST
warmupDurationSecs: 60 # New pod warmup
connectionPool:
tcp:
maxConnections: 200
connectTimeout: 5s
http:
http2MaxRequests: 500
maxRequestsPerConnection: 100
idleTimeout: 300s
outlierDetection:
consecutiveErrors: 5
interval: 30s
baseEjectionTime: 30s
maxEjectionPercent: 50
```
**使用场景：**
* 高性能 REST API
* 响应时间可变的请求
* Pod 性能存在差异的环境
### 示例 2：基于用户会话的路由
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: web-session-affinity
spec:
host: web-frontend
trafficPolicy:
loadBalancer:
consistentHash:
httpCookie:
name: "session-id"
ttl: 7200s # 2 hours
minimumRingSize: 2048
connectionPool:
tcp:
maxConnections: 500
http:
http1MaxPendingRequests: 100
maxRequestsPerConnection: 50
```
**使用场景：**
* Web 应用程序会话持久化
* 购物车一致性
* 每用户缓存利用
### 示例 3：多区域全局服务
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: global-api-multi-region
spec:
host: global-api
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST
localityLbSetting:
enabled: true
distribute:
# US West
- from: us-west-1/*
to:
"us-west-1/*": 80
"us-west-2/*": 15
"us-east-1/*": 5
# US East
- from: us-east-1/*
to:
"us-east-1/*": 80
"us-east-2/*": 15
"us-west-1/*": 5
# EU
- from: eu-central-1/*
to:
"eu-central-1/*": 90
"eu-west-1/*": 10
failover:
- from: us-west-1
to: us-west-2
- from: us-east-1
to: us-east-2
connectionPool:
tcp:
maxConnections: 1000
http:
http2MaxRequests: 2000
outlierDetection:
consecutiveErrors: 5
interval: 30s
baseEjectionTime: 60s
```
**使用场景：**
* 全球 SaaS 服务
* 最小化延迟
* 每区域灾难恢复
* 跨 AZ 流量成本降低
### 示例 4：缓存服务优化
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: cache-service-optimized
spec:
host: redis-cache
trafficPolicy:
loadBalancer:
consistentHash:
httpHeaderName: "x-cache-key"
minimumRingSize: 4096 # Large ring size to minimize redistribution
connectionPool:
tcp:
maxConnections: 100
connectTimeout: 1s
http:
http1MaxPendingRequests: 20
maxRequestsPerConnection: 1000
idleTimeout: 600s
outlierDetection:
consecutiveErrors: 3
interval: 10s
baseEjectionTime: 30s
```
**使用场景：**
* 最大化缓存命中率
* 一致的缓存 key 路由
* 分片策略
### 示例 5：数据库连接池
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: database-connection-pool
spec:
host: postgres-primary
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST
connectionPool:
tcp:
maxConnections: 50 # DB connection limit
connectTimeout: 5s
http:
http1MaxPendingRequests: 10
maxRequestsPerConnection: 100
outlierDetection:
consecutiveErrors: 3
interval: 60s
baseEjectionTime: 120s
```
**使用场景：**
* 数据库连接池管理
* 慢查询分配
* 连接限制强制执行
### 示例 6：高流量处理
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: high-traffic-service
spec:
host: analytics-ingestion
trafficPolicy:
loadBalancer:
simple: RANDOM # Fast selection to minimize overhead
connectionPool:
tcp:
maxConnections: 5000
connectTimeout: 1s
http:
http2MaxRequests: 10000
maxRequestsPerConnection: 1000
idleTimeout: 60s
outlierDetection:
consecutiveErrors: 10
interval: 30s
baseEjectionTime: 30s
maxEjectionPercent: 20 # Limited ejection at scale
```
**使用场景：**
* 事件收集（Analytics）
* 日志收集
* 大规模数据处理
## 算法选择指南
### 决策树
```mermaid
flowchart TD
Start[Load Balancing Needed]
Q1{Session
Persistence
Required?}
Q2{Very High
Traffic?}
Q3{Variable
Response Times?}
Q4{Geographic
Distribution?}
Q5{TCP
Proxy?}
R1[CONSISTENT_HASH
httpCookie or
httpHeaderName]
R2[RANDOM]
R3[LEAST_REQUEST]
R4[LEAST_REQUEST +
Locality Setting]
R5[PASSTHROUGH]
R6[ROUND_ROBIN
Default]
Start --> Q1
Q1 -->|Yes| R1
Q1 -->|No| Q2
Q2 -->|Yes| R2
Q2 -->|No| Q3
Q3 -->|Yes| Q4
Q3 -->|No| Q4
Q4 -->|Yes| R4
Q4 -->|No| Q5
Q5 -->|Yes| R5
Q5 -->|No| R6
%% Style definitions
classDef question fill:#326CE5,stroke:#333,stroke-width:1px,color:white;
classDef result fill:#00C7B7,stroke:#333,stroke-width:1px,color:white;
classDef start fill:#f9f9f9,stroke:#333,stroke-width:1px,color:black;
%% Class applications
class Start start;
class Q1,Q2,Q3,Q4,Q5 question;
class R1,R2,R3,R4,R5,R6 result;
```
### 按服务类型推荐的算法
| 服务类型 | 推荐算法 | 原因 |
| ------------------ | ------------------------- | ------- |
| **REST API** | LEAST\_REQUEST | 响应时间一致性 |
| **GraphQL API** | LEAST\_REQUEST | 复杂查询分配 |
| **gRPC** | LEAST\_REQUEST | 流式负载均衡 |
| **Web 前端** | CONSISTENT\_HASH (cookie) | 会话持久化 |
| **WebSocket** | CONSISTENT\_HASH (header) | 连接持久化 |
| **缓存服务** | CONSISTENT\_HASH (header) | 缓存命中率 |
| **Analytics/日志收集** | RANDOM | 大规模处理 |
| **数据库** | LEAST\_REQUEST | 连接池管理 |
| **静态内容** | ROUND\_ROBIN | 简单且足够 |
| **消息队列** | LEAST\_REQUEST | 队列负载均衡 |
| **批处理** | LEAST\_REQUEST | 作业分配 |
### 按流量模式选择
```yaml
# 1. Uniform small requests (< 10ms)
trafficPolicy:
loadBalancer:
simple: ROUND_ROBIN # Simple and efficient
# 2. Variable requests (10ms ~ 1s+)
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST # Load adaptive
# 3. Very high traffic (10,000+ RPS)
trafficPolicy:
loadBalancer:
simple: RANDOM # Minimize overhead
# 4. Session-based (user state)
trafficPolicy:
loadBalancer:
consistentHash:
httpCookie:
name: "session-id"
ttl: 3600s
# 5. Multi-region
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST
localityLbSetting:
enabled: true
```
## 最佳实践
### 1. 算法选择原则
**良好示例：**
```yaml
# API with variable response times
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: api-service-best-practice
spec:
host: api-service
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST # Load adaptive
warmupDurationSecs: 60 # New pod warmup
```
**不佳示例：**
```yaml
# Using ROUND_ROBIN with variable response times
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: api-service-bad-practice
spec:
host: api-service
trafficPolicy:
loadBalancer:
simple: ROUND_ROBIN # Causes load imbalance
```
### 2. 连接池必需设置
始终将连接池与负载均衡一同配置：
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: complete-lb-config
spec:
host: api-service
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST
connectionPool: # Required
tcp:
maxConnections: 100
http:
http1MaxPendingRequests: 50
maxRequestsPerConnection: 100
```
### 3. 异常值检测组合
与 Circuit Breaker 一起使用，以移除发生故障的 Pod：
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: lb-with-outlier
spec:
host: api-service
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST
outlierDetection: # Required
consecutiveErrors: 5
interval: 30s
baseEjectionTime: 30s
```
### 4. 使用一致性哈希时的注意事项
```yaml
# Good example: Using session storage
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: web-with-redis-session
annotations:
description: "Uses Redis for session storage"
spec:
host: web-frontend
trafficPolicy:
loadBalancer:
consistentHash:
httpCookie:
name: "session-id"
ttl: 3600s
# Using Redis session storage maintains
# sessions even on pod restart
```
```yaml
# Caution: Local session only
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: web-local-session-only
annotations:
warning: "No external session storage - sessions lost on pod restart"
spec:
host: web-frontend
trafficPolicy:
loadBalancer:
consistentHash:
httpCookie:
name: "session-id"
ttl: 3600s
# Warning: Sessions lost on pod restart
```
### 5. 多区域部署
```yaml
# Good example: Locality + Failover
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: multi-region-best-practice
spec:
host: global-service
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST
localityLbSetting:
enabled: true
distribute:
- from: us-west/*
to:
"us-west/*": 80
"us-east/*": 20
failover: # Required
- from: us-west
to: us-east
```
### 6. 监控和指标
监控负载均衡的有效性：
```yaml
# Prometheus queries
# Request distribution per pod
sum by (destination_workload) (rate(istio_requests_total[5m]))
# Response time per pod
histogram_quantile(0.95,
sum by (destination_workload, le) (
rate(istio_request_duration_milliseconds_bucket[5m])
)
)
# Active connection count
sum by (destination_workload) (envoy_cluster_upstream_cx_active)
```
### 7. 逐步应用
```yaml
# Step 1: Default ROUND_ROBIN
simple: ROUND_ROBIN
# Step 2: LEAST_REQUEST after monitoring
simple: LEAST_REQUEST
# Step 3: Add Connection Pool
simple: LEAST_REQUEST
connectionPool: ...
# Step 4: Add Outlier Detection
simple: LEAST_REQUEST
connectionPool: ...
outlierDetection: ...
```
### 8. 文档
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: api-service-lb
annotations:
# Configuration rationale
purpose: "Distribute load based on active requests"
# Algorithm selection basis
algorithm-rationale: |
- LEAST_REQUEST: Response times vary 10ms-500ms
- warmupDurationSecs: New pods need 60s to warm up cache
# Test results
test-results: |
- Load test: 1000 RPS evenly distributed
- P95 latency: 150ms (improved from 300ms with ROUND_ROBIN)
- No pod overload observed
# Monitoring
monitoring: |
- Dashboard: grafana.example.com/d/istio-workload
- Alert: High P95 latency > 500ms
```
## 故障排除
### 负载分配不均衡
**症状：**
```bash
# Check CPU usage per pod
kubectl top pods -n production
# Output:
# NAME CPU MEMORY
# api-pod-1 80% 2Gi
# api-pod-2 20% 1Gi
# api-pod-3 15% 1Gi
```
**原因和解决方案：**
```yaml
# 1. Change ROUND_ROBIN -> LEAST_REQUEST
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: api-service-fix
spec:
host: api-service
trafficPolicy:
loadBalancer:
simple: LEAST_REQUEST # Changed
# 2. Add Warmup
warmupDurationSecs: 60
# 3. Add Outlier Detection
outlierDetection:
consecutiveErrors: 5
interval: 30s
baseEjectionTime: 30s
```
### 一致性哈希不均衡
**症状：**
```bash
# Check Envoy metrics
kubectl exec -it pod-name -c istio-proxy -- \
curl localhost:15000/stats/prometheus | grep upstream_rq_total
# Requests concentrated on specific pod
```
**解决方案：**
```yaml
apiVersion: networking.istio.io/v1
kind: DestinationRule
metadata:
name: hash-fix
spec:
host: api-service
trafficPolicy:
loadBalancer:
consistentHash:
httpHeaderName: "x-user-id"
minimumRingSize: 4096 # Increase 2048 -> 4096
```
### 基于位置的路由不起作用
**验证步骤：**
```bash
# 1. Check pod locality labels
kubectl get pods -o jsonpath='{range .items[*]}{.metadata.name}{"\t"}{.metadata.labels.topology\.kubernetes\.io/region}{"\t"}{.metadata.labels.topology\.kubernetes\.io/zone}{"\n"}{end}'
# 2. Check Istio Proxy configuration
istioctl proxy-config endpoints pod-name | grep locality
# 3. Check DestinationRule application
istioctl proxy-config clusters pod-name --fqdn api-service.default.svc.cluster.local -o json | jq '.[] | .localityLbEndpoints'
```
**修复：**
```yaml
# Check and add node labels
apiVersion: v1
kind: Node
metadata:
labels:
topology.kubernetes.io/region: us-west
topology.kubernetes.io/zone: us-west-1
```
## 参考资料
* [Istio Load Balancing](https://istio.io/latest/docs/reference/config/networking/destination-rule/#LoadBalancerSettings)
* [Envoy Load Balancing](https://www.envoyproxy.io/docs/envoy/latest/intro/arch_overview/upstream/load_balancing/load_balancing)
* [Consistent Hashing](https://www.toptal.com/big-data/consistent-hashing)
* [Locality Load Balancing](https://istio.io/latest/docs/tasks/traffic-management/locality-load-balancing/)