Zabbix optimization, zabbix
Zabbix optimization 1
Recently, zabbix has been sending the following alarms. When it appears, the host obviously feels slow, and 300 + IOPS and 30 + % cpu IOwait can be seen in actual monitoring.
Zabbix housekeeper processes more than 75% busy
This process is used by zabbix to clean up the historical data of the database, because zabbix contains
History storage period
History storage period (in days)
The two parameters are used as the data retention period. When the data in the database exceeds the storage period, the process deletes the data.
In addition, the parameter History storage period (in days) affects the sampling interval of the monitoring server and the data size of the server.
The primary solution to this problem is to optimize the database.
The basic data optimization is as follows:
Check whether the database uses innoDB:
Log on to the zabbix database and run the following command:
SELECT TABLE_NAME, ENGINEFROM information_schema.TABLESWHERE TABLE_SCHEMA = 'zabbix ';
Use as much memory as possible for data. Configure the following parameters as needed:
Innodb_buffer_pool_size =
For InnoDB multi-task running, if the memory is greater than 1 GB, you can configure multi-task running. Configure the number of task instances based on the memory size allocated by the memory. The official recommendation is to allocate a buffer of no less than 1 GB for each instance.
Innodb_buffer_pool_instances =