ZABBIX 5.0: automatically monitor Alibaba cloud RDS through LLD

Blog： Blog Garden personal
I did before RDS monitor , because RDS More examples , Manual addition is not efficient , So it's changed to LLD(Low-level discovery) Way to monitor .

What is? LLD

LLD(Low-level discovery), That is, low-level discovery , It provides a way to automatically create monitoring items for different entities on the computer , Trigger and graphic methods . for example ,Zabbix You can start monitoring the file system or network interface automatically on your machine , Instead of manually creating monitoring items for each file system or network interface . Besides , You can configure the Zabbix According to the actual results after regular implementation of the discovery , To remove unwanted monitoring .

Users can define their own discovery types , As long as they follow specific JSON agreement .

Collect data script

Call alicloud Api, collection RDS related data , For relevant configurations, please refer to Previous post , Collection script omitted .

Need to put Api Processing of returned data , Change the field to {#MACRO} Formal LLD macro , The last generation json Formatted data ：

for example ：

[{
	"{#DBINSTANCEID}": "rr-XXX",
	"{#DBNAME}": "XXX"
}, {
	"{#DBINSTANCEID}": "rm-XXX",
	"{#DBNAME}": "XXX"
}, {
	"{#DBINSTANCEID}": "rm-XXX",
	"{#DBNAME}": "XXX"
}]

explain ：

• {#DBINSTANCEID}：RDS example ID
• {#DBNAME}：RDS Instance name

Tips：LLD Macros must be capitalized , Otherwise, an error will be reported .

The configuration template

New template , Create autodiscover rules ：

explain ：

Parameters	explain
name	Rule name
type	Types of inspections found
Key value	It can be a built-in key value , It can also be customized
Update interval	Time interval between discovery execution
Custom time interval	Non essential election , flexible ： Allows you to redefine the default update interval , Dispatch ： It can make the monitored item effective at a specific time or time series
Insufficient resource cycle	Resource expiration time , If the resource fails, delete , If set to “0”, The entity will be deleted immediately . Not recommended “0”

Create a monitoring prototype ：

Create trigger prototype ：

Create graphic prototypes ：

And so on , Add all the items that need to be monitored .

See the effect ：

summary

LLD The way , It greatly reduces the burden of artificially adding monitoring ,RDS There is no need to add, delete and monitor the changes of instances , Truly achieve automatic monitoring .

But it also brings some problems , For example, performance problems , The more examples , The more prototypes you monitor , The more monitoring and collection script processes executed at the same time , The greater the load pressure on the server , And frequently call Alibaba cloud Api It will also bring some costs .

therefore , This monitoring scheme is limited to Alibaba cloud RDS Scenarios with a small number of instances . If there are many instances , Consider increasing the cache , Read data once , Store data in cache , Then monitor the client to read data from the cache , Reduce the frequency of invoking collection scripts .