All Products
Document Center

TCP health check

Last Updated: Nov 06, 2019

You can create health check for an address pool to monitor each IP address in the pool separately and obtain the running status of the application services corresponding to the IP addresses. When a monitored IP address encounters an exception, the address will be blocked until it is restored.

Health check name

Set a health check name that is easy to recognize. It can contain both Chinese and English characters.

TCP monitoring

Target IP addresses are monitored over TCP. The metrics include network accessibility, port availability, and latency.

Port checking

Target IP addresses are monitored over TCP to check whether the corresponding ports support the telnet action. If telnet is successful, the service is running properly; if telnet fails, a service exception occurs.

Time-out period

The time of responding to sent TCP data packets is calculated during TCP monitoring. Health check is deemed to have timed out when no response packet is returned within the time-out period. The time-out period can be set to 2, 3, 5, or 10, in seconds.

Continuous failed attempts

During TCP monitoring, the system determines that an application service is abnormal only after consecutive exceptions have occurred. This prevents transient jitters and other causes from affecting monitoring accuracy. Continuous Failed Attempts can be set to 1, 2, or 3 times.

  • 1 times: An application service is deemed abnormal when health check detects one alarm.
  • 2 times: An application service is deemed abnormal when health check detects two alarms.
  • 3 times: An application service is deemed abnormal when health check detects three alarms.

Failure Rate

The failure rate is the proportion of abnormal metrics to the total metrics of health check during TCP monitoring. When the failure rate exceeds the configured threshold, the system determines that the application service is abnormal. Failure Rate can be set to 20%, 50%, 80%, or 100%.

Monitored node

A monitored node is the geographic location of the node that executes TCP monitoring. The default monitored nodes are:

Nodes geographic location
BGP Nodes Zhangjiakou、 Qingdao、Hangzhou、Shanghai、Hohhot、Shenzhen、Beijing
Overseas Nodes Hong Kong、Germany、Singapore、Silicon Valley、Australia、Malaysia、Japan
ISP Nodes ChinaUnicomWuhan、ChinaUnicomDalian、ChinaUnicomNanjing、ChinaUnicomTianjin、 ChinaTelecomQingdao、ChinaTelecomChangsha、ChinaTelecomXian、ChinaTelecomZhengzhou、ChinaMobileWuhan、ChinaMobileGuangzhou、ChinaMobileDalian、ChinaMobileNanjing