What are TCP health checks?
TCP is one of the network protocols over which GTM performs health checks on the required addresses. TCP health checks are used to monitor the required IP addresses and collect multiple metrics over TCP. These metrics include the network accessibility, port availability, and latency. If a monitored IP address is abnormal, the system blocks the abnormal IP address. If the IP address becomes normal, the system unblocks the IP address.
Features
1. Health Check Protocol
This parameter specifies a protocol over which GTM monitors the required IP addresses and collect multiple metrics. These metrics include the web server network accessibility, port availability, and latency. We recommend that you select TCP.
2. Check interval
This parameter specifies an interval between health checks during TCP monitoring. Default value: 1 minute. The Ultimate Edition supports a minimum interval of 15 seconds.
3. Select a port
This parameter specifies a port number. When GTM monitors the required IP addresses over TCP, GTM checks whether the port of the required IP addresses allows access over Telnet. If access over Telnet is allowed, the application service that is running on these IP addresses is deemed normal. If it is not, the application service is deemed abnormal.
4. Time Out
This parameter specifies a timeout period. During TCP monitoring, the system calculates the response time of each TCP packet that is sent. If the response time of a packet exceeds the specified timeout period, a health check times out. Valid values: 2 seconds, 3 seconds, 5 seconds, and 10 seconds.
5. Continuous Failed Attempts
During HTTP or HTTPS monitoring, if several exceptions occur, an application service is deemed abnormal. This configuration ensures the accuracy of monitoring. The accuracy may be affected by some issues, such as instantaneous network jitter. Valid values: 1 Times, 2 Times, and 3 Times.
1 Times: An application service is deemed abnormal when the health check feature detects one alert.
2 Times: An application service is deemed abnormal when the health check feature detects two alerts.
3 Times: An application service is deemed abnormal when the health check feature detects three alerts.
6. Failure Rate
During TCP monitoring, the system checks whether a web server functions as expected based on a status code that is returned from the web server. If a return code is greater than the specified threshold, an application service is deemed abnormal. Valid values: 20%, 50%, 80%, and 100%.
7. Monitoring Nodes
This parameter specifies the geographical locations of nodes on which you can perform TCP monitoring. The following table describes the predefined monitoring nodes.
Node type | Geographical locations |
|---|---|
BGP | Zhangjiakou, Qingdao, Hangzhou, Shanghai, Hohhot, Shenzhen, and Beijing. |
Overseas | Hong Kong, Germany, Singapore, Silicon Valley, Australia, Malaysia, and Japan. |
ISP | ChinaUnicomWuhan, ChinaUnicomDalian, ChinaUnicomNanjing, ChinaUnicomTianjin, ChinaTelecomQingdao, ChinaTelecomChangsha, ChinaTelecomXian, ChinaTelecomZhengzhou, ChinaMobileShenzhen, ChinaMobileDalian, and ChinaMobileNanjing. |
Procedure
1 . Log on to the Alibaba Cloud DNS console.
2 . In the left-side navigation pane, click Global Traffic Manager. On the Global Traffic Manager page, click the required instance ID.
If the GTM instance is not configured, you must configure the instance. For more information, see Overview and Enable the health check feature.
If the instance is configured, click the Address Pool Configurations tab. On this tab, turn on the Health Check switch, and click Edit. You can modify the required parameters based on the features of TCP health checks.