Lecture 205: Route 53 - Health Checks Hands On
- 课程：[NEW] AWS Certified Advanced Networking Specialty 2026
- 视频时长：04:39
- 来源：Udemy Lecture 28556411
- Slide 数量：7
- 画面切换：00:50、01:38、02:10、02:43、03:16、03:55
扫描说明
逐条读取了 Udemy 页面中的 125 条英文 Transcript cue，并按时间轴扫描全长约 04:39 的视频。依据 Endpoint 基本配置、Advanced Configuration、其余 Region Check、制造 Network Failure、诊断失败原因、Calculated Check、CloudWatch Alarm 与最终状态划分为 7 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. 为 us-east-1 创建 HTTP Endpoint Health Check
时间范围：00:00–00:50
总结 / Summary
中文
进入 Route 53 Health Checks，为 us-east-1 EC2 创建 Endpoint Check。Monitor 可按 IP Address 或 Domain Name，本例选 IP 54.172.8.44、Protocol HTTP、Port 80、Path /。真实应用常提供 /health 等专用 Endpoint，让 Response 更准确反映 Application Health，而不仅是 Web Server 能否响应。
English
In Route 53 Health Checks, create an endpoint check for the us-east-1 EC2 instance. An endpoint can be specified by IP address or domain name; this demo uses IP 54.172.8.44, HTTP, port 80, and path /. A real application commonly exposes a dedicated path such as /health so the response reflects application health rather than merely web-server reachability.
Transcript
00:00 So let's go ahead and
00:01 create some health checks.
00:02 So on the left hand side,
00:03 I'm going to go into the health checks, and we're going to
00:05 create health checks for all our EC2 instances.
00:08 So we're going to check three health checks.
00:09 So the first one is going to be from an
00:11 instance in US east one, and is going to be an end point.
00:15 And then you'd just specify either an IP address or a
00:18 domain name.
00:19 So we'll keep it as an IP address.
00:21 And my instance in US east one is right here.
00:24 So we'll paste that in. We have to specify a port,
00:27 so we'll keep it as 80, because this is the HTTP port.
00:30 And for path,
00:31 we're just going to be keeping the path as slash because
00:33 while this is the same as the IP slash,
00:36 which is the root of our website,
00:37 if we had a real application,
00:39 sometimes it is very common to have a path slash health,
00:42 for example,
00:42 which responds with the health of the endpoint itself.
00:46 Okay, so we have this ready,
00:48 so I'll just remove the slash health.
2. Advanced Configuration 控制频率、阈值、内容与 Checker Region
时间范围：00:50–01:38
总结 / Summary
中文
Advanced Configuration 可选择 Standard 30 秒或成本更高的 Fast 10 秒 Interval、Failure Threshold、前 5,120 Bytes 的 String Matching、Latency Graph、Invert Status、Disable Check，以及自定义或 Recommended Health Checker Regions。还可创建 Alarm 通知失败；本例保留默认 Standard/Recommended 且暂不通知。
English
Advanced configuration offers a standard 30-second or higher-cost fast 10-second interval, failure threshold, string matching in the first 5,120 bytes, latency graphs, status inversion, disabling, and custom or recommended checker regions. An alarm can also notify on failure. The demo keeps the standard and recommended defaults and creates no notification.
Transcript
00:50 And then we can look at some advanced configuration.
00:52 So we can have either have a standard
00:54 every thirty seconds or a
00:55 fast every ten seconds health check.
00:57 We'll keep it as standard because this is
00:59 otherwise more expensive.
01:01 How many times does it need to fail
01:02 before being considered as a failure?
01:05 Do you want to do string matching so do you want to look for
01:08 a string in the first 5,120 bytes? Yes or no?
01:10 Do we want a latency graph,
01:12 to see how latency evolves over time?
01:17 Do we want to invert the health check status.
01:20 Do we want to be healthy and vise-versa.
01:22 Or disable it.
01:23 And then do you want to customize the regions of
01:25 the health checkers?
01:26 Or do you want to use the recommended and we'll just keep it
01:28 as is with using recommended?
01:30 So every option is pretty much as default.
01:32 And do we want to be notified whenever this health check
01:35 fails? Yes or no by creating alarm?
01:37 I will just say no for now.
3. 按同样方式创建 Singapore 与 Frankfurt Endpoint Check
时间范围：01:38–02:10
总结 / Summary
中文
完成 us-east-1 后，再为 ap-southeast-1 与 eu-central-1 创建同类 HTTP Endpoint Check，各自使用对应 EC2 Public IP 并用 Region 名称命名。这样 Health Check 列表中有三条一一对应的 Regional Check，后续可以直接关联 Route 53 Routing Record。
English
After us-east-1, create equivalent HTTP endpoint checks for ap-southeast-1 and eu-central-1 using their EC2 public IPs and regional names. The health-check list now has three one-to-one regional checks that can later be associated directly with Route 53 routing records.
Transcript
01:38 So we have created our first health check.
01:40 Now let's create our second health check and it's going to
01:43 be for AP Southeast one. So AP Southeast one,
01:48 and then IP address right here,
01:51 not host name. And then next and create.
01:54 And the last health check is for EU central one.
01:58 So let's create this health check and I will name it
02:01 EU central one.
02:04 And then here is the IP address and click on next and
02:08 create health check.
4. 删除 Singapore Security Group 的 HTTP Rule 制造 Failure
时间范围：02:10–02:43
总结 / Summary
中文
为验证 Network Failure Detection，打开 Singapore EC2 的 Security Group Inbound Rules，删除允许 TCP/80 的 IPv4 与 IPv6 HTTP Rule，只保留 SSH。Route 53 Checker 随后无法连接该 Public Endpoint，ap-southeast-1 Check 应在达到 Failure Threshold 后从 Healthy 转为 Unhealthy。
English
To test network failure detection, open the Singapore EC2 security group's inbound rules and remove the IPv4 and IPv6 HTTP rules for TCP port 80, leaving SSH. Route 53 checkers can no longer connect to that public endpoint, so the ap-southeast-1 check should become unhealthy after reaching the failure threshold.
Transcript
02:10 Okay. So our health checks are created.
02:12 And what I'm going to do is I'm going to go to one of my
02:15 instances, for example,
02:15 the one in Singapore and for the security group,
02:18 I'm going to start blocking the port 80, removing this rule.
02:23 And the idea is that I want to get a failing health check.
02:25 So I'm going to go into the security group right here.
02:27 I will do action and end it's the inbound rules,
02:31 and it will delete my HTTP based rules.
02:33 And what this will do is that's the one health check
02:36 for AP Southeast one should give me a unhealthy status.
02:39 So let me wait a little bit for the health checkers to do
02:41 their thing, and then we'll get back to you.
5. Health Check 显示 Unhealthy 与 Connection Timeout 根因
时间范围：02:43–03:16
总结 / Summary
中文
等待 Checker 更新后，Singapore 显示 Unhealthy，而 eu-central-1 与 us-east-1 保持 Healthy。Health Checkers/Last Failed Check 会列出 Checker IP、检查时间与 Failure Reason；本例明确报告 Connection Timed Out，并提示 Endpoint、Internet Connection 或 Firewall 阻断，对应刚删除的 Security Group HTTP Rule。
English
After the checkers update, Singapore is unhealthy while eu-central-1 and us-east-1 remain healthy. Health Checkers and Last Failed Check show checker IPs, timestamps, and failure reasons. Here they report connection timeout and suggest the endpoint, internet connection, or firewall, matching the removed security-group HTTP rule.
Transcript
02:43 Okay. So as we can see, we have three health checkers,
02:45 and one of them is unhealthy,
02:48 obviously because I locked at the security group and the
02:50 other two healthy because I don't see any
02:51 security group event.
02:52 So we can have a look at the health checkers and to give you
02:55 some information around when it was last checked and so on.
02:57 And for the unhealthy one, we can view the error status.
03:00 So if we look at view last failed check,
03:03 we can see that there was a connection timeout,
03:06 and maybe the requests are being blocked by my firewall and
03:09 firewall is your security group. So that makes sense.
03:11 So at least it gives us some information and they are
03:13 working just as expected.
6. Calculated Check 选择三条 Child，并要求全部 Healthy
时间范围：03:16–03:55
总结 / Summary
中文
创建 What to monitor = Status of other health checks 的 Calculated Check，把 ap-southeast-1、us-east-1、eu-central-1 三条 Child 加入。Condition 可设至少 N 条 Healthy、全部 Healthy（AND）或任意一条 Healthy（OR）；本例选择 all health checks are healthy，因此任一 Child 失败都会让 Parent 失败。
English
Create a calculated check using Status of other health checks and add the ap-southeast-1, us-east-1, and eu-central-1 children. The condition can require at least N healthy checks, all healthy checks (AND), or one or more healthy checks (OR). The demo requires all, so any failed child makes the parent fail.
Transcript
03:16 And one last thing and you can create is a calculated
03:18 health check.
03:19 And this one is going to be a calculated health check,
03:23 Cal-cu-lated. Here we go.
03:25 And this going to check for the status of
03:27 other health checks.
03:29 And now we can specify which health checks
03:32 we want to monitor.
03:33 And okay,
03:34 we're saying maybe you want to report healthy when one of
03:37 the three health checks are healthy or when 2,
03:39 or when all of them are healthy.
03:41 So this is an AND or one or more health checks are healthy.
03:44 So this is,
03:45 so we can definitely create as complicated rule.
03:47 So I would just keep it as this should be healthy when all
03:50 of my health checks are healthy and then
03:52 click on next, next.
7. CloudWatch Alarm 可代表 Private Health，Parent 状态按规则汇总
时间范围：03:55–04:39
总结 / Summary
中文
另一种 Check 可监控指定 Region 的 CloudWatch Alarm，从而把 Private EC2 等无法被 Public Checker 直连的 Resource Health 接入 Route 53；演示因没有现成 Alarm 未创建。最终 Calculated Check 显示 Unhealthy，Description 为 threshold 3 of 3，且其 Child 列表是一条 Unhealthy、两条 Healthy，完全符合 AND 定义。
English
Another health check type monitors a CloudWatch alarm in a selected region, letting private EC2 health feed Route 53 without direct public checker access; the demo has no alarm available to create it. The final calculated check is unhealthy with threshold 3 of 3 because its child list contains one unhealthy and two healthy checks, exactly matching the AND rule.
Transcript
03:55 And we have created a calculated health check and the last
03:58 kind of a health check we can create is to monitor the
04:01 state of a CloudWatch alarm.
04:02 In which case we need to specify the region
04:04 the alarm is going to be in,
04:06 and then this alarm could be monitoring.
04:08 Obviously the state have a private EC2 instance
04:10 for example.
04:11 And this is how we would link the health check,
04:13 the health of a private resource
04:15 into a route 53 health check. Okay.
04:17 But I can't create it right now because I don't have
04:19 an alarm available for us. Okay.
04:22 Okay.
04:22 So my calculated health check is now reported unhealthy
04:24 because well, one of the health checks,
04:26 I'm trying to monitor is unhealthy,
04:28 and this is how we define it.
04:29 So that really shows you the power of health checks.
04:31 And in the next lecture, we'll be using them of course,
04:33 alongside records in route 53.
04:35 So I hope you liked this lecture and I will see you in the
04:38 next lecture.
整讲总结 / Overall Summary
中文
Hands-on 创建了三条 Regional HTTP Endpoint Health Check，演示了 Interval、Threshold、String Matching 等选项；随后通过删除 Singapore 的 TCP/80 Security Group Rule 制造 Connection Timeout，并在 Route 53 中诊断 Unhealthy。最后用 3-of-3 Calculated Check 汇总三条 Child，同时说明 Private Resource 可通过 CloudWatch Alarm 接入。
English
The hands-on creates three regional HTTP endpoint health checks and reviews interval, threshold, string matching, and related options. Removing Singapore's TCP/80 security-group rule produces a connection timeout that Route 53 diagnoses as unhealthy. A 3-of-3 calculated check then aggregates the three children, while CloudWatch alarms provide the path for private-resource health.
返回顶部 ↑