健康检查模式架构

Architecture

健康检查模式架构图，展示负载均衡器探针、验证数据库、缓存、磁盘和依赖状态的深度健康检查、自动实例轮换以及与 PagerDuty 集成的连续故障告警。该模板模拟实现自愈系统的健康监控基础设施，不健康的实例自动从轮换中移除并通知运维团队。对于构建具有适当可观测性的生产就绪服务至关重要。

下载 .fz 文件

完整 FlowZap 代码

LoadBalancer { # Load Balancer
n1: circle label:"Health Check Timer"
n2: rectangle label:"Send Health Probe"
n3: rectangle label:"Evaluate Response"
n4: diamond label:"Instance Healthy?"
n5: rectangle label:"Keep in Rotation"
n6: rectangle label:"Remove from Rotation"
n7: rectangle label:"Update Routing Table"
n1.handle(right) -> n2.handle(left) [label="Every 10s"]
n2.handle(bottom) -> ServiceA.n8.handle(top) [label="GET /health"]
n2.handle(bottom) -> ServiceB.n12.handle(top) [label="GET /health"]
n3.handle(right) -> n4.handle(left)
n4.handle(right) -> n5.handle(left) [label="200 OK"]
n4.handle(bottom) -> n6.handle(top) [label="Timeout/5xx"]
n5.handle(right) -> n7.handle(left)
n6.handle(right) -> n7.handle(left)
}
ServiceA { # Service A
n8: rectangle label:"Check Database Connection"
n9: rectangle label:"Check Redis Connection"
n10: rectangle label:"Check Disk Space"
n11: rectangle label:"Return Health Status"
n8.handle(right) -> n9.handle(left) [label="DB OK"]
n9.handle(right) -> n10.handle(left) [label="Cache OK"]
n10.handle(right) -> n11.handle(left) [label="Disk OK"]
n11.handle(top) -> LoadBalancer.n3.handle(bottom) [label="Status JSON"]
}
ServiceB { # Service B
n12: rectangle label:"Check API Dependencies"
n13: rectangle label:"Check Message Queue"
n14: rectangle label:"Check Memory Usage"
n15: rectangle label:"Return Health Status"
n12.handle(right) -> n13.handle(left) [label="APIs OK"]
n13.handle(right) -> n14.handle(left) [label="Queue OK"]
n14.handle(right) -> n15.handle(left) [label="Memory OK"]
n15.handle(top) -> LoadBalancer.n3.handle(bottom) [label="Status JSON"]
}
Alerting { # Alerting System
n16: rectangle label:"Monitor Health Trends"
n17: diamond label:"Consecutive Failures?"
n18: rectangle label:"Send PagerDuty Alert"
n19: rectangle label:"Log Health Metric"
n16.handle(right) -> n17.handle(left)
n17.handle(right) -> n18.handle(left) [label="3+ Failures"]
n17.handle(bottom) -> n19.handle(top) [label="Below Threshold"]
n7.handle(bottom) -> Alerting.n16.handle(top) [label="Status Change"]
}

为什么需要这个工作流？

Load balancers that route traffic to unhealthy instances cause user-facing errors. Deep health checks verify not just that the process is running, but that all critical dependencies (database, cache, disk, external APIs) are functioning—enabling automatic instance rotation and proactive alerting before users are affected.

工作原理

Step 1: The load balancer sends periodic health probes (GET /health) to each service instance.
Step 2: Each service checks its database connection, cache connectivity, disk space, and memory usage.
Step 3: A JSON health status is returned with individual component statuses.
Step 4: Healthy instances remain in the load balancer rotation; unhealthy ones are removed.
Step 5: The alerting system monitors health trends and fires alerts on consecutive failures.
Step 6: PagerDuty notifications are sent when failure thresholds are exceeded.

替代方案

Simple TCP health checks only verify the process is listening. Liveness probes in Kubernetes restart unhealthy pods. Readiness probes control traffic routing. This template shows deep application-level health checks with alerting integration.

Key Facts

Template Name	健康检查模式架构
Category	Architecture
Steps	6 workflow steps
Format	FlowZap Code (.fz file)