Lecture 204: Route 53 - Health Checks
- 课程：[NEW] AWS Certified Advanced Networking Specialty 2026
- 视频时长：04:54
- 来源：Udemy Lecture 28556407
- Slide 数量：5
- 画面切换：01:08、01:35、03:14、04:00
扫描说明
逐条读取了 Udemy 页面中的 133 条英文 Transcript cue，并按时间轴扫描全长约 04:54 的视频。依据 Automated DNS Failover、三类 Health Check、Public Endpoint 检查、Calculated Health Check、Private Resource 监控划分为 5 个小节。截图均在 Transcript 侧栏关闭、播放器控件隐藏后复核，统一为 1364×770。
Transcript 来自 Udemy 页面自带的英文自动字幕，没有使用 ASR。下面保留页面原文，包括其中的语法或识别错误。
1. Health Check 把 Routing Record 变成 Automated DNS Failover
时间范围：00:00–01:08
总结 / Summary
中文
Multi-Region Application 可用 Latency 等 Policy 把 User 指向更合适的 Public ALB，但 Region 故障时必须停止返回失效 Endpoint。Route 53 Health Check 监控各 Public Resource，并与相应 DNS Record 关联；当检查失败时 Route 53 从 Answer 中排除该 Record，从而实现 Automated DNS Failover。
English
A multi-region application can use a policy such as latency routing to direct users to a suitable public ALB, but it must stop returning a failed regional endpoint. Route 53 health checks monitor the public resources and associate with their DNS records. Failed records are then excluded from answers, providing automated DNS failover.
Transcript
00:00 So let's talk about Health Checks
00:01 in Route 53.
00:02 So health checks are a way for you to check
00:05 the health of mainly public resources,
00:07 although there's a way for us to do it
00:09 for private resources as well, as we'll see in this lecture.
00:11 So the idea is that, for example,
00:12 we have two Load balancers in different regions
00:15 and they're public load balancers, okay?
00:17 And behind the scenes,
00:18 we have our application running in both of them.
00:20 So we're running into a multi-region setup
00:22 because we want high availability, and so on,
00:24 at the region level.
00:25 Then we're going to use Route 53 to create DNS records.
00:29 So that's when users access our URL, for example,
00:32 mydomain.com, then they get redirected to, for example,
00:35 the closest load balancer they have.
00:38 So this would be the case with a latency type of record.
00:41 But we want to make sure that, if one region is down,
00:44 then we don't send our users to that region,
00:46 obviously, right?
00:47 So to do so,
00:48 we're going to create health checks from Route 53.
00:50 So we'll create health checks on the one in us-east-1,
00:53 and we will create a health check on our instance
00:56 in eu-west-1.
00:58 Well, with these two health checks,
00:59 we're going to be able to associate them
01:01 with our Route 53 records.
01:04 And the reason we do so is to get automated DNS failover.
2. 三类检查：Endpoint、Calculated 与 CloudWatch Alarm
时间范围：01:08–01:35
总结 / Summary
中文
Route 53 支持三种 Health Check：直接 Monitor Public Endpoint；把多个 Child Check 组合成 Calculated Health Check；或监控 CloudWatch Alarm。第三种给了最灵活的 Metric 来源，例如 DynamoDB Throttle、RDS Alarm 或 Custom Metric，也成为检查 Private Resource 的关键方法。所有 Health Check 自带 CloudWatch Metric 可供观察。
English
Route 53 supports three health-check types: monitoring a public endpoint directly, combining child checks into a calculated health check, or monitoring a CloudWatch alarm. The third option accepts flexible metrics such as DynamoDB throttles, RDS alarms, or custom metrics and is key for private resources. Each health check also publishes a CloudWatch metric.
Transcript
01:08 So we have three health checks that are possible.
01:10 The ones I just showed you, which are the health check
01:12 that monitor an endpoint, which is a public endpoint.
01:14 So it could be an application, a server,
01:16 or another AWS resource.
01:18 It could be a health check
01:18 that monitors other health checks,
01:20 also called a calculated health check,
01:22 or it could be a health check
01:23 that monitors a CloudWatch Alarm,
01:25 which gives you more control and is helpful for private
01:27 resources as we'll see in this lecture.
01:30 Finally, these health checks have their own metric
01:32 and you can view them in CloudWatch metrics as well.
3. 约 15 个 Global Checker 按协议、阈值与响应内容判定 Endpoint
时间范围：01:35–03:14
总结 / Summary
中文
Endpoint Health Check 由约 15 个全球 Checker 发起，支持 HTTP、HTTPS、TCP；默认每 30 秒检查，也可付更高成本使用 10 秒 Fast Health Check。可配置 Healthy/Unhealthy Threshold；超过 18% 的 Checker 报告 Healthy 时 Route 53 判为 Healthy。HTTP(S) 要返回 2xx/3xx，还可在 Response 前 5,120 Bytes 中查找指定 Text。Network 上必须允许来自 Route 53 Health Checker IP Range 的 Inbound Request。
English
About 15 global checkers probe an endpoint over HTTP, HTTPS, or TCP. The normal interval is 30 seconds; a higher-cost fast check runs every 10 seconds. Healthy and unhealthy thresholds are configurable, and Route 53 considers the endpoint healthy when more than 18% of checkers report healthy. HTTP(S) requires a 2xx or 3xx response and can search the first 5,120 response bytes for text. Network rules must allow inbound requests from Route 53 health-checker IP ranges.
Transcript
01:35 So let's look at how health checks work
01:37 with a specific endpoint.
01:38 So if we have a health check for eu-west-1, for an ALB,
01:41 then the health checkers of AWS
01:44 are coming from all around the world.
01:45 So it's not just one health checker.
01:47 It's about 15 health checkers from all around the world.
01:49 And they're all going to send requests
01:51 into our public endpoint to wherever routes we set.
01:55 And then if it gets 200 OK code back or the code we defined,
01:58 then the resource is deemed healthy.
02:01 So about 15 global health checkers
02:02 will check the endpoint health,
02:04 and then you can set a threshold for healthy or unhealthy.
02:07 You can set an interval,
02:08 so we have two options.
02:09 It could be either 30 seconds for regular health checks
02:12 or every 10 seconds, which is a higher cost,
02:14 which is what's called a fast health check.
02:16 It supports many protocols, so HTTP, and HTTPS, and TCP.
02:20 And the rule is that if over 18% of the health checkers
02:24 say that the endpoint is healthy,
02:26 then Route 53 will consider it healthy,
02:28 otherwise it's deemed unhealthy.
02:30 And you have the ability to choose
02:31 which locations you want to use for the health checks.
02:34 Now the health checks will only pass if you have the status
02:36 2xx or 3xx status code back from the load balancer
02:40 and the health check has a cool capability.
02:42 So if it is a text-based response,
02:45 then the health checkers can check the first 5,120 bytes
02:50 of the response to look for some specific texts
02:52 in the response itself.
02:53 Finally, very important from a network perspective,
02:56 if you want for it to work, obviously,
02:58 the health checkers must be able to access your
03:01 Application Balancer or whatever endpoints you have.
03:04 And so therefore you must allow incoming requests
03:06 coming from the Route 53 health checkers' IP address range.
03:09 And you can find this address range at the URL
03:12 in the bottom right of the screen.
4. Calculated Health Check 用 OR、AND、NOT 汇总 Child 状态
时间范围：03:14–04:00
总结 / Summary
中文
Calculated Health Check 把多个 Child Health Check 的结果汇总为一个 Parent，可使用 OR、AND、NOT 逻辑，最多监控 256 个 Child，并规定至少多少个 Child 必须通过。它适合表达 Composite Availability 条件，也可配合人为控制的 Child 在 Maintenance Window 中改变整体状态，而不必让所有底层检查真的失败。
English
A calculated health check combines child health-check results into one parent using OR, AND, or NOT logic. It can monitor up to 256 child checks and define how many must pass. This supports composite availability rules and can include a manually controlled child to change the parent state during maintenance without forcing all underlying checks to fail.
Transcript
03:14 Now the second type of health checks we have
03:16 are calculated health checks.
03:18 And so this is to combine the results
03:20 of multiple health checks into a single health check.
03:22 And so if you look at Route 53,
03:24 with three EC2 instance,
03:25 we can create three health checks.
03:27 They're all going to be children health check,
03:28 and they can all monitor each EC2 instance one by one.
03:31 And then we can define a parent health check,
03:33 which is going to be defined
03:35 on all these child health checks.
03:38 And so the conditions to combine all these health checks
03:40 could be an OR, an AND, or a NOT.
03:43 You can monitor up to 256 child health checks,
03:47 and you can specify how many of the health checks
03:49 need to pass to make the parent pass.
03:51 So the use case for this,
03:53 for example, if you want to have
03:54 a parent health check to perform maintenance
03:56 on your website without causing
03:58 all the health checks to fail.
5. Private Endpoint 不能被 Public Checker 直连，改用 CloudWatch Alarm
时间范围：04:00–04:54
总结 / Summary
中文
Route 53 Checker 位于 Public Web，无法直接访问 Private VPC Endpoint 或 On-premises Private Resource。解决方法是在内部建立反映 Resource Health 的 CloudWatch Metric，再创建 CloudWatch Alarm，并让 Route 53 Health Check 监控该 Alarm。Alarm 进入 ALARM State 时 Health Check 变为 Unhealthy，即可把 Private Health 间接纳入 DNS Failover。
English
Route 53 checkers live on the public web and cannot directly reach private VPC endpoints or private on-premises resources. Instead, publish an internal CloudWatch metric, create an alarm from it, and configure a Route 53 health check to monitor that alarm. When the alarm enters ALARM state, the health check becomes unhealthy, bringing private-resource health into DNS failover indirectly.
Transcript
04:00 And so how do we monitor the health of a private resource?
04:03 So in case you want to monitor something private,
04:06 it's going to be difficult because
04:08 while all the Route 53 health checkers
04:09 live on the public web, they're outside of your VPC,
04:12 so they cannot access private endpoints.
04:14 So if it's a private VPC or an on-premises resource.
04:18 And so the way we can do it, though,
04:19 is to create a CloudWatch Metric
04:21 and assign a CloudWatch Alarm on it.
04:24 And then you can assign the CloudWatch Alarm
04:25 into the health checker.
04:27 So the idea is that we're going to monitor
04:28 the health of our EC2 instance in a private subnet
04:31 with a CloudWatch Metric.
04:32 And then if the metric is breached,
04:34 we're going to create a CloudWatch Alarm on it.
04:37 And when the alarm goes into the alarm state,
04:39 then the health checker is going to be
04:41 automatically unhealthy
04:43 and therefore will have created exactly what we want,
04:45 which is a health check on a private resource,
04:48 which is the most common use case on how to do it.
04:50 So that's it for this lecture.
04:51 I hope you liked it and I will see you in the next lecture.
整讲总结 / Overall Summary
中文
Route 53 Health Check 将 Resource Health 与 DNS Record 关联，实现自动 DNS Failover。Public Endpoint 可由全球 Checker 直接按 HTTP/HTTPS/TCP、阈值和内容检查；复杂条件用 Calculated Health Check；Private Resource 则通过 CloudWatch Metric/Alarm 间接监控。务必允许 Health Checker IP Range 访问 Public Endpoint。
English
Route 53 health checks connect resource health to DNS records for automated failover. Global checkers can directly probe public endpoints using HTTP, HTTPS, or TCP plus thresholds and content matching; calculated health checks express composite conditions; and CloudWatch metrics and alarms provide indirect monitoring for private resources. Public endpoints must allow the health-checker IP ranges.
返回顶部 ↑