Project develops a fork of the Zabbix monitoring system aimed at increasing efficiency, performance, and scalability, and suitable for creating fault-tolerant configurations dynamically deployed across multiple servers. Initially, the project as a set of patches to enhance the performance of Zabbix, but in April, work began on creating a standalone fork. The project's code is licensed under GPLv2.
Under high loads, Zabbix users encounter the lack of clustering as such in the free version, and issues when large volumes of data need to be stored in the DBMS. The relational DBMS supported in Zabbix, such as PostgreSQL, MySQL, Oracle, and SQLite, are poorly suited for trend storage in history — sampling a large number of metrics over six months will already be 'heavy,' necessitating optimization of the DBMS and queries, as well as building clusters of database servers, etc.
As a solution, Glaber embodies the idea of using a specialized DBMS , which provides good data compression and very high request processing speeds (on the same hardware, a reduction in load on CPU and disk systems by 20-50 times can be achieved). In addition to ClickHouse support, Glaber also has various optimizations, such as the use of asynchronous SNMP requests, batch processing of data from monitoring agents, and the use of nmap for parallel host availability checks, which allowed the states polling to be accelerated by more than 100 times. Glaber is also working on supporting , for which future plans include utilizing .
Source: opennet.ru
