All Products
Search
Document Center

Enable health check

Last Updated: Mar 26, 2021

Overview

If you enable the health check feature, you can perform health checks for the addresses in an address pool to monitor the availability of your application service in real time. This way, automatic fault isolation and switchover is achieved. Available health check options include ping, tcp, and http(s).

Prerequisites

A GTM instance is created, an access policy is set, and an address pool is created.

Procedure

1. Add a Health Check item

On the Address Pool Configurations page, click the + icon in front of the address pool name, and then click Add to configure the health check feature.

1

2 . Configure the health check feature

You must configure the health check feature for the addresses in an address pool to obtain the availability of your application service in real time. This way, automatic fault isolation and switchover is achieved. You can set the following parameters:

  • Check Protocol: the type of the protocol. Valid values: ping, tcp, and http(s). Default value: ping.

  • Check Interval: the interval at which the monitoring node performs health checks. Default value: 1 minute.

  • Number of Ping Packets: the total number of ping packets that are sent to the address for each check. This metric is available if the Check Protocol parameter is set to ping. Default value: 20.

  • Packet Loss Rate: the packet loss rate. The packet loss rate in a check = The number of lost packets/The total number of ping packets × 100%. This metric is available if the Check Protocol parameter is set to ping. Default value: 10%.

  • Timeout Period: the timeout period of a response. If a response is not sent within the timeout period, an alert is triggered. This metric is available if the Check Protocol parameter is set to ping. Default value: 5 seconds.

  • Number of Consecutive Failures: specifies whether to retry if an error occurs during a health check. Default value: 2.

  • Failure Ratio: the failure ratio of health checks. Each monitoring node independently performs health checks at the same time. The failure ratio = The number of nodes that have alerts/The total number of monitoring nodes × 100%. This metric is available if the Check Protocol parameter is set to ping. Default value: 50%.

  • Monitoring Node: the location of the monitoring node. This metric is available if the Check Protocol parameter is set to ping.

  • Enable SNI: If SNI is enabled, the hostname is sent to the monitoring node during a TLS handshake.

Notice

If the address of the application service is located in mainland China, do not select a monitoring node in a region outside mainland China. If the address of the application service is located in a region outside mainland China, do not select a monitoring node in mainland China. This way, you can prevent false alerts that are triggered due to network problems.

3 . Monitor nodes

The health check feature supports the following monitoring nodes:

2